The Design of Switching Circuits

Citation preview

WILLIAM KEISTER, ALISTAIR E. RITCHIE, SETHH.WASHBURN MEMBERSOF THE TECHNICALSTAFF BELLTELEPHONELABORATORIES, INC.

.

.

'

.

· • •A broad account of the principles and design methods , ' .' • •.·of ·switching circuits as usfkl for auto.rnatic control. .~.·

,.:

•

~-- ~---

•

.

• ..

·.·•.,·

~

• •

:

.

·""

_ .. ;_·

.....

·.~..

....

•

•

• • =~-- •

.

•

•

~

•

$8.00

T

HIS BOOK presents the basic techniques of switching circuit designtechniques which are applicable to simple control circuits as well as to telephone switching systems, digital computers, and other large control systems. These techniques extend today throughout a wide range of applications in science and technology, in business and in industrywhere switching circuits are used to perform a variety of complicated control actions and mathematical calculations. This book covers specifically the methods of using interconnected two-valued devices, primarily electromagnetic relays, in systems for automatic control. The treatment includes a brief description of circuit components; fundamental principles of relay contact network design, including switching (Boolean) algebra; design of basic circuits to perform specific functions such as counting, memory, translating, lockout, and selecting; and the integration of such functional circuits into systems. Also included is the application of switching algebra to electronic circuit design. The techniques are applicable lo the design of digital computers, automatic telephone switching systems, and other systems where response is a function of the combination and sequence of discrete input conditions. The approach is based upon the logical interrelation of dose.d or open circuit paths rather than upon electric circuit theory.

More than 400 illustrations of circuits, of time and sequence charts and of many other topics clarify the treatment. Many problems demonstrate the important phases of circuit design .

...

~rurg

·µr.·t~umcn·i1i

I~

~

The Design of Switching

Circuits

THE BELL TELEPHONE LABORATORIES

SERIES

Speech and Hearing - Harvey Fletcher Probability and Its Engineering Uses - Thornton C. Fry Elementary Differential

Equations (Second Edition) - Thornton

- K. S. Johnson

Transmission Circuits for Telephonic Communication Transmission Networks and Wave Filters

- T. E. Shea

Economic Control of Quality of Manufactured Product Electromechanical Transducers and Wave Filters

C. Fry

- W. A. Shewhart

(Second Edition)

- Warren

P. Mason

Electromagnetic Waves - S. A. Schelkunoff Network Analysis and Feedback Amplifier

Design - Hendrik

W. Bode

Servomechanisms - LeRoy A. MacColl Quartz Crystals for Electrical

Circuits

Capacitors: Their Use In Electronic Fourier Integrals for Practical Applications

- R. A. Helsing

Circuits

- M. Brotherton

- George A. Campbell

and Ronald M. Foster

Visible Speech - Ralph K. Potter, George A. Kopp, and Harriet Applied Mathematics for Engineers and Scientists Earth Conduction Effects In Transmission

C. Green

- S. A. Schelkunoff

Systems - Erling

D. Sunde

Radar Systems and Components - Members of the Staff of the Laboratories Introduction by M. 1. Kelly Theory and Design of Electron Beams - 1. R. Pierce Piezoelectric Crystals and Their Application to Ultrasonics Microwave Electronics

The-Design of Switching Circuits

- George C. Southworth

Tubes - 1. R. Pierce

Electrons and Holes in Semiconductors Ferromagnetism

P. Mason

- John C. Sla~er

Principles and Applications of Waveguide Transmission Traveling-Wave

- Warren

- William

Shockley

- Richard M. Bozarth

- William Keister,

AlisWr

E. Ritchie.

and Seth H. Washburn

THE DESIGN OF SWITCHING CIRCUITS BY

WILLIAM KEISTER ALISTAIR E. RITCHIE SETH H. WASHBURN MEMBERS OF THE TECHNICAL STAFF BELL TELEPHONE LABORATORIES

Third Printing

D. VAN NOSTRAND COMPANY, INC. TORONTO

NEW YORK

LONDON

NEW YORK

D. Van Nostrand Company, Inc., 250 Fourth Avenue, New York 3 TORONTO

D. Van Nostrand Company (Canada), Ltd., 228 Bloor Street,

Toronto

LONDON

Macmillan & Company, Ltd., St. Martin's Street, London, W.C. 2

Copyright, 1951, by D. VAN NOSTRAND COMPANY, INC. Published simultaneously in Canada by D. Van Nostrand Company (Canada), Ltd.

All Rights Reserved This book, or any parts thereof, may not be reproduced In any form without written permission from the authors and the publishers. Fint Published September 1961 Reprinted February 1962, January 1963

Printed in the United States of America

To N., J., and J.

FOREWORD The present is an excellent time for this book to appear. In the past, general interest in the switching art and knowledge of its unique techniques were limited to a few quarters where complex control mechanisms such as telephone switching systems were developed and used. Now, however, the situation is changing rapidly and radically. College professors, research engineers, scientists, and mathematicians are now aware of and keenly interested in the subject. This change was brought about, or at least greatly accelerated, by the appearance of automatic digital computing systems. Fundamentally, such systems employ essentially the same techniques as those useful in telephone switching. However, their ability to perform complicated mathematical calculations furnished a brilliant and readily understandable demonstration of the fact that switching circuits can be designed to accomplish, mechanically, at least certain sorts of operations which are normally associated with human mental effort. For example, they can follow trains of orders, they can make use of tables, they can make deductions following formal logical rules. Of course, in other respects such mechanisms may be sadly lacking in comparison with human intelligences. The abilities they do have, however, seem great enough to promise the elimination of much of the mental drudgery required by our modern world. It is this glimpse of a coming age of "mechanized intelligence" that recently caught the imagination of professional people and caused the upsurge of interest in switching. If this textbook succeeds in nourishing that interest, and if it helps its readers to work toward the ultimate realization of such a significant goal, the effort that went into its preparation will indeed have served a very worthy purpose. The book is a natural product of the Bell Telephone Laboratories, since the design of switching circuits to establish telephone connections is a basic and vital activity of this organization. When we make a telephone call, we spin our dial a few times, and - in a matter of seconds - the particular phone that we want among millions of others is ringing. This happens with such regularity and such reliability that we have come to take it for granted. Most of us never try, or for that matter are not able, to visualize the series of intricate events that take place automatically in a modern dial telephone office before that called phone is rung. - vii -

viii

The Design of Switching Circuits

In simple terms, what happens is that at the instant we lift _our handset, a group of switching circuits springs obedient!! to our service. These switching circuits identify our line as the one m need of attention, connect themselves to it, give us the signal to proceed, count our dial pulses, memorize the series of digits, determine our call's destination locate and test the available paths, select the most preferred ' . . combination, join the links end-to-end, and fmally, when the connection is all set, ring the wanted phone. All this is cuits of the dial noting that while ditional groups, random requests

accomplished for us in seconds by the switching circentral office. Just for perspective, it is also worth one group of switching circuits works on our call, adin perfect teamwork, are taking prompt care of the of thousands of other customers~

Obviously, then, switching circuits used in dial central offices are swift and versatile mechanical servants; and even if they did no more than establish connections, when we are once aware of their existence, we are likely to be intrigued by them. However, they go further in their performance. To facilitate proper charging for their services, these switching circuits also record who calls whom, and when, and how long each conversation lasts. Such competence on their part certainly increases our respect for them. Even this is not all. As a little extra service during their daily work, these central office circuits make a second attempt at setting up a connection in case their first attempt fails; or they locate alternative routes to a destination if the preferred one is not available, or furnish maintenance men with precise information regarding troubles, and so on. This roll-call of events in a dial central office conveys, perhaps, an adequate sense of the type of actions that switching circuits can perform. It also reveals how far the ingenuity, imagination, and precise logic of design engineers have advanced the art in a particular field of application. Such a revelation may induce readers of this textbook to go beyond its Foreword. Those who do, will find a wealth of switching fundamentals, succinctly rationalized and presented by the authors. Since the material for this book was prepared as part of a postwar training program for young engineers at the Bell Telephone Laboratories, it is appropriate to close this Foreword with a simple tribute to the Laboratories' management, particularly Mr. A. B. Clark, Vice President, for initiating and firmly supporting such a training program on the basis of its long-term merits, when there was intense

Foreword

ix

need for manpower in more immediately profitable switching development projects. In years to come, the training program will bring a harvest in the form of technical contributions by the young engineers whose minds have been alerted to the challenges of the switching art. For the present, it is hoped that this book, a direct outcome of the training program, will be a tangible source of satisfaction to those whose official support it represents. JOHN MESZAR

July 1951

PREFACE

This book is not a text on telephone systems. It is concerned, rather, with the basic techniques of switching circuit design: techniques which are applicable to digital' computers and other complex control systems as well as telephone switching systems. The writing of this text was started soon after the end of World War II as a series of lecture notes for use in training new engineers in the Switching Systems Development Department of the Bell Telephone Laboratories. During the succeeding years the notes were revised and the methods of instruction were improved. In 1950, with general interest in the subject of switching increasing, arrangements were made with the Massachusetts Institute of Technology to give to graduate students a onesemester course in switching circuit design. The text for this course was based on the material used for training within the Laboratories. The present volume is a final edition of the text used in the M. I. T. course, revised to take advantage of the academic experience gained there. The objective of this book is to present the fundamental principles underlying the design of switching circuits rather than to describe the operation of specific circuits, except as examples. The material is planned to be complete within itself, no previous knowledge of switching being assumed. For this reason explanations are often given in considerable detail. The book is suitable for individual study as well as for a one-semester or two-semester course of formal instruction. The only prerequisites for its use are a logical mind and a knowledge of elementary electrical circuit theory. The subjects of the chapters range from elementary concepts to the design of circuits containing a relatively large number of switching devices. Chapter 1 serves as an introduction to the subject. Chapter 2 discusses the physical and electrical characteristics of the various devices used for switching. Chapters 3 through 10 give the fundamentals of switching circuit design, including the logical concepts involved and the form of Boolean algebra which has been given the name "Switching Algebra". Chapters 11 through 21 discuss "unifunctional circuits", which are aggregates of switching devices performing single specific switching actions. These are the building blocks of large switching systems. The last three chapters contain a brief treatment of the methods of synthesizing such basic functional circuits into practical

- xi -

xii

The Design of Switching Circuits

working systems. Principles and techniques are illustrated primarily by relay switching circuits. However, electronic techniques employing tubes, rectifiers, and transistors are discussed and many examples are given. Of particular interest in this regard are Chapter 10, which discusses electronic circuit fundamentals, and Section 21. 7 of Chapter 21, which illustrates the application of switching algebra to the design of electronic switching circuits. On the basis of teaching experience, it is felt that design practice is the best way to gain facility in the application of principles. Following most chapters is a set of suitable problems which have been tested by actual classroom use. Since almost all switching problems have a number of satisfactory answers, the solutions are sometimes difficult to judge. As a guide, the approximate quantity of apparatus required for a good solution is given in many of the problems. This is not always the minimum possible apparatus but represents a reasonable point to which a circuit may be reduced. In learning the subject, it is also helpful for the student actually to build circuits from original designs. Many of the problems are suitable for this activity. Particular attention is called to the special problem given at the end of Chapter 11. The basic ideas presented in this text have, for the most part, been distilled from the accumulated work of practicing circuit design engineers. A number of people have contributed directly to the work leading up to the publication of this book. Mr. A. E. Joel was associated with the authors in the early stages of the teaching program and contributed valuable ideas, particularly in the organization of functional circuits. Messrs. G. R. Frost and W. S. Hayward prepared the first drafts of several chapters. Messrs. R. W. Roberts and H. N. Seckler assisted in the details of producing the final text. The book was designed by Mr. K. M. Collins, who not only handled all problems of publication and printing, but made available to the authors a wealth of information and experience in the fine art of preparing a book for public presentation. Finally, the authors wish to express their appreciation for the friendly advice, encouragement, and discerning criticism of Mr. John Meszar in all of the activities leading up to the production of this book. WILLIAM KEISTER ALISTAIR E. RITCHIE SETH H. WASHBURN

luly 1951

LIST

OF CHAPTERS

CHAPTER

AND MAJOR

1

Fundamental

CHAPTER

2.1 2.2 2.3 2.4 2.5

PAGE 1

Control

Concepts

2

Switching

PAGE 9

Apparatus

General- Purpose Relays ............................ Special-Purpose Relays ............................. Electronic Devices and Electromagnetic Switches ........... Miscellaneous Switching Apparatus ..... - ................ Battery and Ground Conventions on Circuit Drawings .........

CHAPTER

4

4.8

4.9 4.10

23 24 26

. . . .

Page Page Page Page

27 30 31 34

. . . . . . . . . . .

Page Page Page Page Page Page Page Page Page Page Page

36 37 38 40 43 45 46 50 54 55 64

Configurations

Logic of Contact Networks ........................... "And-Or" and Series-Parallel Relationships ............... The Negative Relation .............................. Basic Simplifications In Two-Terminal Network Design Bridge-Type Networks ............................. Multi-Terminal Networks ........................... Transfer Contacts In Networks ........................ Transfer-Tree Circuits ............................. Transfer Chains .................................. Reiterative Networks .............................. Problems ......................................

- xiii

10

2l

PAGE 36

Relay Contact Network 4.1 4.2 4.3 4.4 4.5 4.6 4.7

Page Page Page Page Page

Paths

Basic Control Methods for Relays with Single Windings ........ Summary of the Four Basic Control Methods ............... Relays with Special Characteristics ..................... Problems ...... .- ...............................

CHAPTER

. . . . .

PAGE 27

3

Basic Relay Control 3 .J 3.2 3.3

SECTIONS

-

.......

The Design of Switching Circuits

xiv CHAPTER 5

PAGE 68

Switching Algebra, and Manipulation of Relay Contact Networks 5.1 5.2 5.3

5.4 5.5

Symbolic Notatior in Switching Algebra ................... The Postulates of the Algebra . ........................

Theorems ....... .............................. General Notes on Network Simplification and Manipulation . ..... Multi-Terminal Networks ...........................

5.6

Tables of Combinations .............................

5.7

Delta-Y Transformations ............................ Resolution and Synthesis of Bridge-Type Networks . ..........

5.8 5.9 5.10

. . •• . . . .

.

Consideration of Basic Relay Operating Path ............... . Recapitulation of the Postulates and Theorems : . . . . . . . . . . . . . References ..................................... . Problems ...................................... .

CHAPTER 8

Page 69 Page 70 Page 71 Page 81 Page 83 Page 89 Page 93 Page 95 Page 99 Page 101 Page 103 Page 103

PAGE 108

Design of Combinational Relay Circuits 6.1 8.2 6.3

6.4 6.5 6.6 6.7

General Design Procedure ........................... Input and Output Conditions .......................... General Types of Relay Circuits ....................... Capabilities of Combinational Circuits ................... Design Procedure for Combinational Circuits .............. Examples of Combinational Circuit Design ................ Temporary False Output Conditions. . . . . . . . . . . . . . .. . . . . . . Problems .•.....................................

CHAPTER 7

. . . . . . .

Page Page Page Page Page Page Page Page

108 109 110 110 111 112 127 128

Page Page Page Page

132 136 137 137

Page Page Page Page Page Page

144

PAGE 131

Time Charts, Sequence Diagrams, Simplified Schematics, and Graphic Descriptions of Relay Circuits 7.1 7.2 7.3

7.4

Sequence Diagrams and Tables . . . . . . . . . . . . . . . . . . . . . . ... Time Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Partial Schematic Drawings or "Shorts" .................. . Graphic Description of Relay Circuits ................... .

CHAPTER 8

PAGE 142

Design of Sequential Relay Circuits 8.1 8.2 8.3

8.4 8.5 8.8 8.7

Control of Secondary Relays. . . . . . . . . . . . . . . . . . . . . ..... . Effect of Race Conditions on Internal Control .............. . Providing Output Conditions in Sequential Circuits ........... . Developing an Operating Sequence ...................... . Planning Secondary Relay Actions ...................... . Illustrative Examples of Sequential Circuit Design ........... . Appendix: Development of an Operating Sequence • for a Group of Relays Problems ••••...................................

155 156 156 159 165

Page 172 Page 175

List of Chapters

and Major Sections CHAPTER

Elementary 9.1 9.2 9.3

9.4 9.5 9.6 9.7 9.8

Rotary Selector Switches: Mechanical Considerations ........ Control of Rotary Selector Switches ..................... Switch Stepping Circuits ............................ Switch RuMing Circuits ............................. Switch Halting Circuits ............................. Switch Restoring Circuits ........................... Crossbar Switches: Mechanical Considerations ............. Control of the Crossbar Switch ........................ Problems ......................................

Elementary

10.2 10.3

11.4 11.5

11.6

12.3

12.4 12.5 12.6

Page Page Page Page Page

183 183 186 189 193

Page 194 Page 196 Page 199

Switching Circuits

11

Page Page Page Page

203 207 232 233

Page Page Page Page Page Page Page Page

236 239 239 242 247 267 272 273

Page Page Page Page Page Page

277 278 280 281 288 289

PAGE 235

for Counting

Systems of Enumeration ............................ . Basis of Relay Counting-Circuit Design .................. . Single-Pulse Counters .............................. . Two-Pulse Counters or Pulse-Frequency Dividers ........... . Counting a Train of Pulses. . . . . . . . . . . . . . . . . . . . . . . . . . . . Sequence Circuits ................................. . Problems ....................................... . Special Problem .................................. .

12

PAGE 275

Codes and Translating 12.2

Page 179

PAGE 202

of Electron Tubes and Semi-Conductors as Switching Elements Basic Tube and Semi-Conductor Circuits for Switching Applications Electrostatic and Magnetic Effects in Electronic Switching Circuits Problems ...................................... .

CHAPTER

12.1

. . . . . . . . .

Characteristics

Circuits 11.2 11.3

10

Electronic

CHAPTER

11.1

PAGE 179

Switch Control Circuits

CHAPTER

10.1

9

xv

Circuits

How Information is Represented ....................... Simple Combinational Codes .......................... Self-Checking Codes ..................... •· ......... The Theory of Error-Detecting and Error-Correcting Codes .... Translating Circuits ............................... Design Techniques for Translating Networks. . . . . . . . . . . . . . .

. . . . . .

The Design of Switching Circuits

xvi

CHAPTER 12 (CONTINUED)

12.7 12.8 12.9 12.10 12.11 12.12

Examples Illustrating Design Techniques ................. Further Examples Based on Standard Codes ............... Symmetric Checking Networks ............... •......... Intermediate Translation ............................ Translation Between Numbering Systems ................. Large Translators ................................ • Problems ....... : ..............................

CHAPTER 13

. . . . . . .

Page Page Page Page Page Page Page

292 295

Page Page • Page Page Page Page Page Page Page Page

307 309 310 312 314 317 318 322 323 325

298 299

300 302 304

PAGE 306

Circuits for Selecting 13.1 13.2 13',3 13.4 13.5 13.6 13.7 13.8

13.9

Elementary Relay Selecting Circuits .................... Multlbranch Trees: General ......................... Theoretical Considerations of Multibranch Trees ............ Practical Considerations: Size of Relays ................. Practical Considerations: Control Codes ................. Modification of Selecting Trees ........................ Use of Selecting Circuits In Large Translators ............. Selection by Sequential Control ........................ Selection with Non-Symmetrical Varistors ................ Problems ......................................

CHAPTER 14

. . . . . . . . . .

PAGE 327

Circuits for Connecting 14.1 14.2 14.3

14.4 14.5 14.6

The Multiple .................................... Many-to-Many CoMectlons by Multiple .................. Links ......................................... Multistage Connecting .............................. Control of Connecting Circuits ........................ Combined Selecting and Connecting ..................... Problems ......................................

CHAPTER 15

. . . . . . .

Page 330 Page 331 Page 333 Page 335 Page 338 Page 342 Page 343

PAGE 344

Circuits for Lockout 15.1

15.2 15.3 15.4

15.5 15.6

The Double-Transfer Lockout Circuit ................... End-Relay Lockout Circuits .......................... Gate-Type Lockout Circuits .......................... Varying the Order of Preference In Lockout Circuits Multistage Lockout Circuits .......................... Electronic Lockout Circuits .......................... Problems ......................................

.........

. . . . . . .

Page Page Page Page Page Page Page

347 351 353 355

357 358 363

List of Chapters

and Major Sections CHAPTER

Circuits 16.1 16.2 16.3 16.4

16

Basic Requirements of Finding and Hunting Circuits Finding Circuits .................................. Hunting Circuits .................................. Electronic Hunting Circuits ......................... Problems ......................................

17

Circuits

Circuits

366 370 370 375 377

. . . . .

Page Page Page Page Page

379 381 383 395 404

. . . . . .

Page Page Page Page Page Page

405 407 414 419 421 422

. . . . .

Page Page Page Page Page

425 427 432 437 443

. . .

Page Page Page Page Page Page Page

446 448 449 451 454 456 459

•..

for Timing

18

. .....

PAGE 405

19

: .....

PAGE 423

for Checking

Affirmative and Negative Check Principles ................ Probability-Type Checking ........................... Checking Single Relays and Contacts .................... Checks for Opens, Grounds, and Crosses ................. Problems ......................................

CHAPTER

Circuits

.

Page Page Page Page Page

PAGE 378

Simple Relay Pulsing Circuits ........................ Capacitor-Timed Relay Pulse Generators ................. Gas-Tube Pulsing Circuits ........................... Vacuum-Tube Pulsing Circuits ........................ Pulse Detection ............................ Problem .......................................

Circuits

20.1 20.2 20.3 20.4 20.5 20.6

. . .

for Pulse Generation

CHAPTER

19.1 19.2 19.3 19.4

..........

Standard Slow-Acting Relays ......................... Interrupter Timing Circuits. . . . . . . . . . . . . . . . . . . Capacitor-Timed Relays ............................ Gas-Tube Timing Circuits ........................... Problems ......................................

CHAPTER

18.1 18.2 18.3 18 .4 18.5

PAGE 365

for Finding and Hunting

CHAPTER

17 .1 17.2 17 .3 17.4

xvii

20

PAGE 446

for Registration

Types of Register Elements .......................... Applications of Register Circuits In Switching Systems ........ Relays as Register Elements ......................... Register Input Control ........................... Register Output Control ............................. Electronic Registration Methods ....................... Problems ......................................

. . . .

The Design of Switching Circuits

xviii CHAPTER 21

PAGE 460

Circuits for Calculating 21.1 21.2 21.3 21.4 21.5 21.6 21.7 21.8 21.9

Codes .............................. •• •• •••• • • • · · , . · · , . · · · · · · Addition ••.••.•.••••..•....•..... Subtraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Multiplication ................................... Division ....................................... Square Roots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Design of an Electronic Binary Adder .................... Appendix I: N2 as a Sum of Factors .................... Appendix II: Determination of Square Roots by Repeated Subtracuon Problems ......................................

CHAPTER 22

• · • . . . . .

Page Page Page Page Page Page Page Page

.

Page 493 Page 495

461 462 472 474 483 485 487 493

PAGE 496

General Principles of Multifunctional Circuit Design 22.1 22.2 22.3

Circuit Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Underlying Factors ................................ . The Four Stages of Design ........... • ................ .

CHAPTER 23

Page 497 Page 498 Page 500

PAGE 506

Planning a Multifunctional Circuit 23.1 23.2 23.3

23.4

Statement of the Problem. . . . . . . . . . . . . . . . . . . . . . . . . . . . Detailed Requirements . _.. . . . . . . . . . . . . . . . . . . . . . . . . . . Preliminary Planning .............................. Developing the Block Diagram . . . . . . . . . . . . . . . . . . . . . . . .

CHAPTER 24

. . . .

Page Page Page Page

506 508 509 515

. . . .

Page Page Page Page Page Page

524 527 531 532 539 545

PAGE 523

Detailed Design of a Multifunctional Circuit 24.1

Pulse Repeating, Pulse-Train Detecting,

24.2 24.3 24.4 24.5 24.6

and Supervisory The Counting Circuit. .............................. The Register Circuit. .............................. The Sequence-Connecting Circuit ...................... The Minor Functions ............................... Conclusion ••••••................................

INDEX

PAGE 547

Holding Circuits

The Design of Switching Circuits

Chapter FUNDAMENTAL

1

CONTROL

CONCEPTS

From a technological point of view, one of the most striking aspects of the present period of human endeavor is the emergence of machines that handle information automatically. In more graphic but less accurate terms, these are machines that "think". With every passing day, not only scientists and engineers but the general public as well are becoming more aware of complex devices or systems that can accept information in one form or another, correlate and digest this information, and arrive at logical conclusions. Mechanisms of this nature are broadly classified as "automatic control systems." Automatic control systems break down into two distinctly different subdivisions: those that function on the basis of continuous input data, or data that vary smoothly from one value to another; and those that function on the basis of data that vary in discrete steps. The subject of this book is the design of circuits for systems of the latter class, properly known as "switching" systems. Among the more spectacular and well-known examples of such systems are the modern large-scale digital computers and the automatic telephone switching systems. Both of these systems accept information in numerical form and produce output results that are precisely related to the input data. The nature of a switching system and its component circuits can perhaps be more clearly illustrated than defined. Modern dial telephone systems are an example. The machinery of a central office contains a network of switching devices by means of which subscriber lines and trunks can be interconnected. Centralized control circuits are provided to control the actions necessary in establishing a connection. These control circuits may range in size from less than a dozen switching elements to a thousand or more. The system is controlled by instructions from the subscriber which consist of a series of pulses generated when a number is dialed. These signals start an intricate train of reaotions which are determined in part by the information received from the subscriber, in part by information collected within the system concerning busy and idle path conditions, and finally by built-in knowledge. This knowledge is placed there by the designer who analyzed a large number of control situations, determined suitable plans of action, and developed switching circuits to carry out these plans.

- 1-

2

The Design of Switching Circuits

A variety of actions are performed. The dial pulses are counted as they arrive and the information is stored as numerical digits in coded form, to be used as required. Various circuit units are interconnected with each other to cooperate in the control. Trunks and interconnecting linkages are selected in groups and tested for busy and idle conditions. On its own initiative a control circuit can pick an idle path from a large group or, if all are busy, can choose an alternate group or make other disposition of the call, such as returning a busy signal. Centralized control circuits can operate any switch in the office used for interconnecting subscriber lines and trunks, and can locate suitable switches for a given call. After a control circuit has completed its action, it will restore itself to a normal condition and be available for use on another call. During this entire process, information, either received as input data or held as stored data, is recorded, translated, manipulated, transmitted, etc., as discrete digital items. For any set of inputs, there is a corresponding logical and invariable output condition or group of alternative output conditions (for instance, a connection to a busy-tone source instead of the called line, if the line is busy). Other examples can be cited to illustrate the varying facets of switching systems, ranging from relatively simple machine controls up through such complex mechanisms as automatic accounting systems and inventory control systems. The latter type of system, for instance, can store information on the complete inventory of a business enterprise, be it a department store, factory, warehouse, or airline. Each inventory item is assigned an identifying code, and with it is associated the present status or quantity of the item. The system can answer requests from key equipments at local or remote points as to the current state of any stock item, and can change information automatically, by subtraction in accordance with filled orders or by addition in accordance with cancellations or new purchases. In all these switching systems, the internal circuitry that relates input to output consists of paths interconnecting discrete-valued apparatus elements. Well-known examples of this type of apparatus are electromagnetic relays, electromagnetic switches, rectifiers, gas-filled electronic tubes, magnetic tapes and drums, mercury delay-lines, and certain arrangements of vacuum tubes. The action of any of these elements in switching circuits is to open or close, or switch, interconnecting paths, or to establish definite conditions over them, in predetermined logical patterns which are ordered by the input information. The outputs generated by a system as a whole may consist of pure information, as a printed page of figures in a computer; control signals, as the motor on-off signals in an elevator system; or a particular

Fundamental

Control

state of the system, telephone office.

Concepts

3

as the establishment

of a new talking

channel in a

The concept of logic is the most noteworthy aspect of switching circuits; in fact, the schematic representation of a relay circuit is a symbolic expression of a set of conditions and their logical conclusion. For example, the configuration of a circuit is largely determined by logical relationships of the form: output X should be closed when inputs A and B occur simultaneously; output Y should be closed when input A occurs before input B; output Z should be closed when either input A or input B occurs. Methods of determining logical relationships from statements of requirements, and the developing of circuits to conform to the logic, form a principal part of this volume.• Furthermore, it will be seen that the terminology and procedures of formal logic are of great assistance in designing circuits. From this point of view, it is not difficult to see that there is some validity to the feeling that machines designed on such a basis "think." Note, however, that the thought processes are in all cases built into the machine and that the machine at most merely reproduces the original plans of the designer. The preceding discussion field of application for switching trol situations subject to logical discrete or numerical in nature, suitable detecting and converting

indicates that there is a vast ultimate circuits, for not only are many contreatment, but they are inherently or can be transposed into this form by devices.

The ability of switching circuits to perform complex as well as simple functions by means of relatively simple apparatus elements is based upon the combinations into which the apparatus elements of the circuits can be manipulated. Any individual combination of several elements can be made to give a unique indication by suitable interconnections. If apparatus elements that can assume any one of five states or "values" are used, one element can indicate five conditions; two elements, twenty-five conditions; three elements, one hundred and twenty-five conditions; etc. If apparatus elements with only two values are used, the same effects are evident. For example, a group of twenty such two-valued elements can be set into more than a million unique combinations.

It can be seen almost intuitively that the simplest and most elemental control conditions are two-valued, that is, they are "on" or "off", "positive" or "negative", "open" or "closed". Indeed, in a sense, it is straining for complication to go beyond two values for control purposes,

•

An understanding of the electrical basis of circuits this phase of the subject is not stressed.

is assumed,

for the most part,

and

4

The Design of Switching Circuits

since marginal conditions and tolerance limits increasingly difficult to meet are thereby introduced. Since it is natural to match the apparatus to the most effective control conditions, the use of two-valued apparatus for switching applications is clearly indicated. Even with this restricted number of v·alues or states per apparatus element, the rate at which the distinct combinations of elements increases with the number of elements permits the handling of the most complex requirements with a relatively small quantity of apparatus. Multi-valued apparatus, such as a multi-terminal electromagnetic switch, is useful in certain special applications in switching circuits. Although there may be a multiplicity of paths through the device, individually the paths are either opened or closed during operation and the apparatus may be analyzed essentially on a two-valued basis. Control of the device is also usually two-valued, consisting either of a repetition of signals over one lead or combinations of on-off signals over several leads. The multi-valued device can normally be considered as an aggregation of two-valued devices, constructed in a package for reasons of economy or convenience. Since it is designed for specialized appliinherent in two-valued cations, it does not offer the circuit flexibility apparatus. Therefore the major effort in this volume is devoted to the design of circuits employing two-valued elements. An important characteristic of switching circuitry is that it can be made to hold or retain a given state after the excitation that produced the state has passed. This is the factor of memory which permits switching circuits to respond to a sequence or time-order of input intelligence, and eliminates the necessity of an input path to each apparatus element in a circuit. An example of the utility of this concept is furnished in the association of a telephone switching office and a subscriber by means of a pair of wires. Over this path, by a sequence of identical signals, the subscriber can control the connection of his line with any one of millions of other subscribers.

The remarks up to now have been general and do not give a clear picture, in a detailed sense, of what switching circuits do. And in this introduction it is not essential to point out the many simple control functions, where effect is obviously related to cause, that circuits of this type can be designed to handle effectively. However, large circuits or systems, involving dozens or hundreds of apparatus elements and performing intricate combinations of control functions, deserve further consideration. Circuits of this type are inevitably composed of many interrelated sub-circuits, each of which performs one or a small number of individual functions. These sub-circuits can be considered as the true building blocks of major switching circuits. Design of a large circuit or a system, then, consists of determining from the requirements

Fundamental

Control

Concepts

5

of the problem the basic and individual functions which are necessitated by circuit logic; designing circuits to perform these basic functions; and finally, integrating or coordinating the functional circuits into the complete system. In the final system, information received from the outside over input channels is recorded and interpreted, decisions are made as to what functional circuits should be called into play, internal information is passed back and forth between circuits, checks are made as to accuracy and validity of performance,and finally output signals or actions appear. Surprisingly, even in a large and complete switching system the majority of switching circuit requirements can be met by a relatively small number of types of functional circuit. Within each type,of course, there are many variations, some due to detailed requirements and some to the whim of the designer. The names of the types of circuit give an immediate clue to the kind of action each performs. Only a few examples are mentioned here; later chapters describe most of the types of circuit in detail. A simple basic circuit is the "register" circuit which records and holds information until required by another functional circuit. The particular configuration of the register circuit depends upon the form or code in which the information is made available. Often associated with the register circuit is the "translating" circuit which converts information from one code to a different code. It is often convenient to receive or record information in a particular code, whereas a different form is required for use or transmittal to other parts of the system. Information is often sent on a repetitive pulse basis, and to receive it a "counting" circuit is necessary. The counting circuit is actuated by successive pulses and can be arranged to count either a definite number of pulses or to give a distinct indication corresponding to each individual pulse. This circuit is also of use in keeping track of a series of events. In most multi-functional circuits, there is need for "connecting" circuits to associate functional parts for short periods or to connect to circuit inputs and outputs. There are many types of connecting circuits, with the differences dependent upon the number of connectable circuits, the duration of individual connections, and the number of leads required. Often associated with connecting circuits are the "locating" circuits which are used to choose one item out of many. Classified under locating circuits are "selecting" circuits, which pick one item from many on a predetermined basis; "hunting" circuits, which pick any available one from several identical items; and "finding"circuits, which identify a particular item that requires service.

6

The Design of Switching

Circuits

The "lock-out" circuit is used to regulate the flow of traffic among the parts of a complex system. Its usual function is to permit but one out of several competing items to act or to have access to a common part of the system. Functional circuits that have less general application than the preceding ones are the "calculating" circuits. These are circuits which perform the basic arithmetical functions of adding, subtracting, multiplying, and dividing. Their. chief application is in calculating or computing systems. This partial list of basic functional switching circuits serves to give an idea of the types of action of which switching circuits are capable in control applications. As mentioned, these circuits have little utility individually, but are the functional building blocks from which complete systems are constructed. It has already been stated that two-valued apparatus elements are at present considered the best devices for general use in automatic control circuits of the switching type. Where multi-valued elements are used, it is in specific applications where economy or efficiency can be achieved by, in effect, combining several two-valued elements into a single piece of apparatus. Among the two-valued elements, the leading contenders for use in existing switching circuits are electromagnetic relays, hot- or cold-cathode gas tubes, and vacuum tubes connected to act as two-valued devices. Other discrete-valued devices which are appearing in switching circuits at an accelerating rate are varistors, transistors, and magnetic recording devices.

Each of these devices has specific characteristics which peculiarly suit it to certain applications in the switching field. An instance of this appears in the requirement for extreme speed, in the microsecond range, which sometimes arises. This inevitably drives the designer to the use of vacuum tubes or the semiconductors. However, as a general tool for the majority of switching applications, the electromagnetic relay has as yet no peer.

Of major importance in this respect is the manner in which the relay indicates its state. This is by means of contacts that may be used in circuit paths which are completely independent, electrically, of the relay operating path. These contacts may be open or closed when the relay is unoperated, and will be in the inverse state when the relay is operated. By placing the contacts of several relays in series, any conclusion requiring a coincidence of conditions can be indicated. By connecting contacts in parallel, a conclusion dependent upon the existence of any one or more conditions out of many can be indicated. In logic, these two alternative sets of conditions would be known as "and" and

Fundamental

Control Concepts

7

"or" requirements. By suitable interconnections, any conceivable combination of "and" - "or" relationships can be expressed with a relay contact network. Of equal importance with this concept is the fact that use of normally open and normally closed contacts permits specifying not only the inclusion, but also the exclusion, of dependent conditions. These considerations make the relay an excellent instrumentality for expressing the logical relationships which are the very basis of switching problems. A secondary factor, which has considerable practical importance, is the fundamental simplicity of the relay as a circuit element. Although the interconnecting paths of a multi-relay circuit viewed as a whole may become highly complex, the control network for each individual relay forms a simple open or closed path which rarely requires additional apparatus for correct functioning. Physically, the relay is easy to comprehend, even though a great deal of knowledge and ingenuity has entered into its design. From the circuit point of view, in most applications the relay is simply a device which opens or closes a multiplicity of paths when power is applied to it, and performs the inverse operation on the paths when power is removed. The importance of this simplicity should not be underestimated. It enables the circuit designer to concentrate on the logical aspects of any set of circuit requirements rather than be diverted by essentially extraneous problems of making the circuit elements perform in a particular manner. A natural consequence of this is that the designer has greater freedom in dealing with the logical aspects of a switching problem when working with relays than with any other type of circuit element. This is a general statement, and does not exclude the choice of other apparatus elements when their characteristics match the requirements of a particular problem. The relay in a general sense has other attributes which contribute to its utility and flexibility as a circuit tool. Its operating time, depending upon construction, ranges from a minimum of a few milliseconds up to about one second, with more extreme values available in special relays. It has seen such extensive use over many years that it is today a highly-developed and reliable device. Of utmost importance is the fact that it is widely available in a great variety of types that offer almost complete freedom in the choice of the correct relay for each particular application. In view of these considerations, the major portion of this volume is devoted to the design of typical relay circuits. There is less loss of generality in this treatment than might appear, however, since the same fundamental logic is applicable to all types of two-valued switching apparatus. The differences in approach to electronic circuits, for instance, arise in the application of detailed design techniques rather than in

8

The Design of Switching Circuits

general principles. In order to indicate the nature of these differences, selected examples of electronic switching design techniques are incorporated in the text, and tube circuits performing functions comparable to relay circuits are presented. Some basic circuits for control of multi-terminal switches are also included for the. same reason. In presenting the subject of this volume, the objective is to progress from the basic design concepts up through the design of relatively complex multifunctional circuits. Although no formal division is made, the text can be considered as falling into three parts: basic relay circuit design fundamentals; the design of single-function circuits; and the planning and design of multifunctional circuits.

Chapter 2 SWITCHING APPARATUS It has been noted in Chapter 1 that the design of switching circuits consists primarily of analyzing the functional circuit requirements in terms of two-valued conditions, and devising logical interrelationships of circuit paths and apparatus to meet the conditions. The choice of apparatus to use in the circuit is in a sense a separate problem which can often be met after the basic circuit design has been completed. This is particularly true with relay circuits, and to a lesser extent with circuits using electronic apparatus. Since this volume is primarily concerned with the fundamentals of relay circuit design, little detailed study of apparatus is included. However, this chapter will discuss the basic characteristics of relays as circuit elements in order that the circuit design chapters need not be interrupted by digressions on the physical aspects of apparatus. Also, the schematic conventions and the vocabulary used in relay circuit design will be given. On the other hand, it will be more convenient to give the comparable description of switches and electronic circuit elements in the chapters devoted to their circuit use. This will be done since the circuit design using such apparatus is more intimately related to the apparatus than is true with relays. In addition, this chapter will mention some types of apparatus such as keys, plugs, and jacks, which are useful adjuncts to circuits. In all cases descriptions will be general since detailed information can be found in manufacturers' specifications and catalogues. A relay is essentially a remotely controlled switch which can open or close contacts when suitable electrical conditions are met. From a circuit point of view, relays can be divided into two broad classes, known as general-purpose relays and special-purpose relays. The general-purpose relay, often called a telephone-type relay, is the work-horse in relay switching circuits around which is constructed the major portion of switching circuitry. In its usual form it is a directcurrent relay operating in a voltage range of a few volts to approximately 100 volts•. In order to be a flexible tool, the general-purpose relay provides a basic structure on which can be placed a variety of • The most common voltage in telephone central offices, for instance, ls 48 volts.

- 9 -

The Design of Switching Circuits

10

'-t-----,,.-WI

RING TERMINALS

ARMATURE

A

UNOPERATED (CONTACTS OPEN)

8

OPERATED (CONTACTS CLOSED)

Fig. 2-1 Essential Features of a Relay

contact combinations and a variety of winding combinations to meet different operating conditions. The contacts in general carry only moderate currents (1.0 ampere or less), primarily for controlling other relays. The acting times of these relays range from a few milliseconds to several hundred milliseconds. The special-purpose relay, on the other hand, is usually designed for a particular type of application, and will vary in construction and .appearance depending upon its function. Its attributes may be extreme speed, high sensitivity, polarized operation, operation on alternating current, contacts of very heavy load-carrying capacity, etc.

2.1

GENERAL-PURPOSE

RELAYS

All electromagnetic relays are composed of the same three basic parts: a magnetic structure' or framework, a winding, and a set of contacts mounted on springs. The winding and the magnetic structure convert electric power into mechanical action; and the set of contacts, driven by the resultant motion, open and close external circuit paths. The general relation of parts is shown on Fig. 2-1.

Switching Apparatus

11

The magnetic structure consists of a core, one or two polepieces, and a movable armature hinged to the rear pole-piece, often called the heel-piece. All these parts are made of soft iron or other magnetic material, and form a closed magnetic circuit except for an air gap between the free end of the armature and either a front polepiece or the core itself. When the relay is energized, the pull is concentrated in the air gap, and the armature moves to close the gap. The relay winding ( or windings) is placed on the core of the structure, and the winding ends are brought out to connecting terminals. The pull developed at the armatur~ is a function of the current, the number of winding turns, the air gap, and the magnetic material and structure. The relay springs carrying the contacts are usually clamped to the structure at one end and are free to move at the other. There are two sets of springs, one set relatively stiff and fixed in position, the other set relatively flexible and linked to the armature through insulating studs or a comb arrangement. Precious -metal contacts are attached to the mating surfaces of the springs. When the armature moves, the contacts either open or close, depending upon their initial condition. Although all general-purpose relays conform in these fundamental aspects, the construction and appearance of different relays vary in accordance with the manufacturer's design objectives and concepts of efficiency and performance. Two representative types of generalpurpose relays are shown on Fig. 2-2. In service, they mount in horizontal positions on flat metal plates, usually with the moving parts in a vertical plane to avoid the effects of gravity. In considering view, the important

the general purpose relay from the circuit point of aspects of the device are: the contact or spring

Fig. 2-2 Typical General-Purpose

Relays

The Design of Switching Circuits

12 -

_!

_L

OR

t

T

OR

f

--1 t TRANSFER

MAKE OR FRONT NORMALLY OPEN S. P.S.T. NORMALLY OPEN

BREAK OR BACK NORMALLY CLOSED S.P. S.T. NORMALLY CLOSED

BREAK-MAKE S.P. D.T.

A

B

C

CONTINUITYTRANSFER MAKE-BREAK

0

Fig. 2-3 Basic Relay Spring Combinations

combinations; the winding combinations; the power requirements; the acting speeds; and the contact current load capabilities. The magnetic structure of the relay is of primary interest only in an indirect sense, because of its effect upon the power and speed capabilities. However, modifications of the basic relay structure to affect its speed characteristics are important. These various aspects of the relay will be taken up in the following paragraphs. Spring Combinations. The two basic relay spring combinations are the "make" and the "break" combinations. These are illustrated in Figs. 2-3A and 2-3B in two standard schematic forms. In this text, the first form shown, which is the communications standard, will be used exclusively. On the figure, alternate acceptable designations are also given in addition to make and break. Two other very common spring combinations, the "transfer" and the "continuity-transfer" or "continuity", are shown as Figs. 2-3C and 2-3D. The transfer normally has the property of opening the break contact before closing the make contact, whereas the continuity insures a make - before - break contact sequence.

-T

JLL

-lio

•

♦

MAKE-MAKE

MAKE- BREAK- MAKE

BREAK-MAKEBREAI(- MAKE

B

C

A

-tc

•

SEQUENCE MAKE PRELIMINARY MAKE BEFORE MAKE

SEQUENCE MAKE BEFORE BREAK PRELIMINARY MAKE BEFORE BREAK

D

E

Fig. 2-4 Speciallzed Spring Combinations

Switching Apparatus

13

Examples of mo1·e specialized spring combinations are shown on Fig. 2-4. In general, these are less universally used than the combinations of Fig. 2-3. Several spring combinations may be grouped on an individual relay in one or more assemblies. The maximum allowable number of individual springs varies with the type of general purpose relay, but usually does not much exceed 24. A large number of different spring arrangements utilizing the basic combinations are, of course, possible. A typical example in 716 two separate spring assemblies is shown ♦ 4 5 associated with the schematic representation of a relay core on Fig. 2-5. The individual springs are often numbered in 2 RELAY accordance with either the order of wirCORE ing terminals in back or the order of springs viewed from the front. 3'12 4 Relays are also available with a 6 !175 large number, of the order of 50 or 60, of single-type springs, such as make contacts. These are somewhat out of the catFig. 2-5 Spring Combinations Associated with Relay Core egory of general purpose relays, and are often designated multi-contact relays.

~+~

Windings and Winding Arrangements. General-purpose relays are available with a variety of different windings and combinations of two or more separate windings. The effect of the number of turns and the resistance of particular windings will be discussed in the section on operating characteristics; in this section, only the various winding arrangements will be covered. The most common winding arrangement is the single winding shown on Fig. 2-6A. The semi-circular dot on the winding symbol indicates the inner end, or the end of the winding closest to the core. Following the single winding in utility is the double winding, shown on Fig. 2-6B, in which the two windings are often designated primary and secondary. In this and other multi-winding cases, the dots can be interpreted to give the directional sense of the windings in order to simplify the making of aiding or opposing connections. Double windings are prepared in a variety of ways, of which the simplest is to place one winding directly on top of the other. This permits the utmost flexibility in providing windings completely independent in characteristics. Other methods of construction may be employed when it is necessary to

The Design of Switching Circuits

14

---r

ORP 1 OR P1

~

A

~ ~

B

1JE

INDUCTIVE -WINDING -NON-INDUCTIVE WINDING

E

C

D

~

F

F

G

Fig. 2-6 Relay Winding Arrangements

achieve a high degree of balance between the two windings. An example of this is the sandwich winding, where part of one winding is laid on the core, the second winding is next added, and finally the remainder of the first winding is completed. Another example is the parallel type where the wires of the two windings are laid on concurrently. In these cases, the designations P 1 and P 2 are sometimes applied to the two windings instead of P and S. Other winding arrangements of a much more specialized nature are shown on the remainder of Fig. 2-6. The arrangements shown in Figs. 2-6E, 2-6F, and 2-6G include resistors wound non-inductively on the cores. These serve no purpose that could not be achieved by external resistors, but it may be convenient and economical to include them in the relay structure. The specialized arrangements have little general utility, but may be of use in particular applications. _ Relay Operating Characteristics and Adjustments. The force which pulls the relay armature to actuate the contact springs is de veloped by current flowing through the winding. Since the winding is inductive, the initial current does not immediately jump to a value determined by the relay resistance and circuit voltage, but rises exponentially at a rate determined by the timt! constant of the winding and certain other electric and magnetic properties of the structure. Thus, the contacts do not move until the current has risen to a level where the magnetic flux generated across the armature air gap provides sufficient force to overcome the initial back tension of the armature. When the armature starts to move, the back tension increases as the contact springs are flexed to a greater and greater extent. At the same time, the operating force increases as a function both of the continuing

Switching Apparatus

15

rise in current and the decreasing reluctance in the magnetic circuit as the air gap closes. The operating force must increase more rapidly than the armature back tension for the relay to operate completely. For a particular type of relay, the force or pull developed across the armature air gap is a function of the product of the current and the number of turns of the winding. This product is generally known as ampere-turns. It can be seen that, for a specific required pull, there must be a balance among the winding turns, the winding resistance, and the voltage level for which the relay is designed. The resistance and the voltage level are related to the normal power-handling capacity of the relay, which, for general-purpose relays, is usually in the range of one to ten watts. When the circuit to the relay winding is opened, current flow ceases rather abruptly, unless the winding is shunted, and the relay starts to release. The decay of magnetic flux is affected by eddy currents in the core, by the magnetic material of the relay, by the residual air gap left when the relay is fully operated, and by any conducting paths encircling the core. All these factors, plus the time for mechanical motion, determine the release time of the relay. A set of mechanical adjustments is specified for any relay. These have to do with unoperated armature air gap, residual armature air gap (in some cases), direction and limits of spring tension, contact pressure, etc. As determined by the tolerance range of these mechanical adjustments, there are a set of four basic electrical requirements that can be applied to a relay and which determine its operation in acircuit. These are the operate, hold, non-operate, and release requirements, as defined below. The principal mechanical variable which can be changed to make the relay meet the requirements is the spring tension against the relay armature. The requirements are normally specified in terms of current flow. Operate Requirement: The current level at which the relay will definitely operate. The current at which the relay can operate may extend well below this level. Hold Requirement: The level to which the current may be reduced, after the relay has operated, which will insure that the relay remains fully operated. Non-Operate Requirement: The current level at which an unoperated relay definitely will not operate. The relay may operate at any current level above the non-operate value. Release Requirement: The level to which the current through an operated relay can be reduced with assurance that the relay will completely return to an unoperated condition.

The Design of Switching Circuits

16

}RELAY RELAY WILL HOLD

--OPR.

•

CURRENT THROUGH RELAY WINDING

RELAY MAY OR MAY NOT HOLD IN THIS RANGE

WILL OPERATE

HOLD

RELAY MAY OR MAY NOT OPERATE IN THIS RANGE

NON·OPR

RLS

RELAY WILL NOT OPERATE

RELAY Willi RELEASE 0

----

---

Fig. 2-7 Relationship among Relay Adjustments

Since the same mechanical condition must permit the relay to meet all four electrical requirements, or all that are specified in a particular case, the electrical requirements are necessarily closely interrelated. Setting any one requirement value, such as operate, automatically establishes limits within which the others must fall. If a relay is adjustable, the spring tension is regulated to a compromise value which permits meeting all specified electrical requirements. The status of the relay is then as shown on Fig. 2-7. In practice, how ever, it is seldom that all four requirements are placed upon a relay. For normal straightforward operation in open and closed circuits, the operate requirement alone may be specified. The others are necessary only to meet marginal conditions, or to hold the relay to special speed require men ts. In order for a relay to function satisfactorily in a circuit, some margin must be left between the circuit current under adverse conditions and the corresponding relay current requirement. Otherwise, a slight change in the mechanical condition caused by temperature or humidity changes or by wear may cause the relay to fail. A margin of 10% is usually considered adequate to take care of this situation. In practice, the worst circuit current is computed under extreme conditions of voltage and resistance limits, and a check made that the resultant value exceeds the operate or hold requirement by 10% or more. For release, the worst circuit current should be less than the

Switching Apparatus

17

release relay requirement by 10% or more, and should be less than the non-operate requirement by a greater amount if possible. Operating and Release Times. General-purpose relays are obtainable with acting times ranging from less than five milliseconds up to approximately a half-second. Although acting times above the order of fifty milliseconds require special methods of construction which will be discussed in a later section, the remarks here will apply to relays in general. For a given mechanical structure, neglecting the effect of any special time-increasing features, the operating time of a relay is a function of several variables. Of these, the most important for present purposes are the spring load on the armature, the unoperated armature air gap, and the power absorbed by the relay winding. For a given winding, the operating time will increase as the spring load on the armature increases and as the unoperated armature air gap increases. The air gap is fixed to a great extent by the spring combination and is not, strictly speaking, an independent variable. For a given spring load, the operating time will decrease as the power to the winding increases. Therefore, the higher the circuit voltage and, assuming the use of welldesigned coils, the lower the coil resistance, the faster will be the relay. These facts are important to the circuit designer since, where speed of operation is important, he must make a careful choice of the winding coil and must not place a large number of contact springs on the relay. The choice of coil is influenced by the circuit voltage, which is generally set by other considerations, and the maximum power that can be applied to the relay without causing excessive heating. The release time of a relay, again neglecting the effect of any special time-increasing features, is primarily a function of operated spring load and operated armature air gap, the time of release increasing as either of these factors decreases in magnitude. It is worth noting that, to a certain extent, the factors which aid in gaining low operate time tend to increase the release time, and vice versa. It is difficult to attain a high degree of precision in the operate and release times of general-purpose relays. These times are affected both by the structural variations in the relays and by electrical variations in the circuit. The result of this is that individual relays of a particular type, meeting all mechanical and electrical adjustment limits, may differ in operate or release times by as much as two or three to one. The designer must constantly keep this in mind when attempting to meet delicate circuit time requirements. An aspect of the time characteristics of relays is the phenomenon of contact stagger. Since a relay does not operate or release instantaneously, there is a finite armature trans it time during which contacts

18

The Design of Switching Circuits

may open or close more or less in a random sequence. Thus, several milliseconds may elapse between the closure of one make contact on a relay and another make contact in the same spring pile-up. The magnitude of the stagger usually increases with the acting time of the relay. Although definite contact sequences can be imposed, either by use of certain spring combinations or by adjustment, it is generally preferable to design the circuit in such a way that it is not affected by random stagger. Relay Construction for Slow Action. The acting time of a relay, either on operate or release, can be increased considerably by any means which delays the build-up of magnetic flux (operate) or retards its decay (release). This can be done most simply by encircling the relay core by a conducting path. When the circuit is closed to the winding, induced currents in this path will generate magnetic flux to oppose the operating flux. When the circuit to the winding is opened, inducec currents will generate flux in the core in a direction which tends to hold the relay operated. The conducting path has no effect when the steady state of the relay has been reached. The most frequently used method of achieving these effects is to place a highly conducting sleeve, usually of copper, over the core. The sleeve may extend over the length of the core, or may be concentrated at either end of the structure. With end sleeves, operating time is increased to a much greater extent when the sleeve is located at the front end of the relay, close to the armature air gap, than when the sleeve is displaced toward the rear of the relay. Release time is affected to a much lesser degree by sleeve position. The effectiveness of sleeves is much greater in increasing the release time than the operate time of a relay. The maximum reliable operate time that can be achieved with a general-purpose relay is well under 0.1 second, while the maximum release time may run as high as 0.5 second or more. It is also true that release time can be more closely controlled than operate time in manufacture. As a result, it is generally preferable to use the release of a relay for delayed action in circuits. Aside from the use of sleeves, the acting time of relays can be secondary increased to some extent by circuit means. A short-circuited winding, for example, will increase the time of operation and release, particularly the latter. The release of a relay by short-circuiting the operating winding will increase the release time appreciably. Finally, as indicated in the preceding section, the operate time of a relay becomes greater as the circuit current is allowed to approach the operate adjustment value.

Switching Apparatus

19

When slow-acting relays, either operate, release, or both, are used in a circuit, it is customary to indicate the nature of the relay on the schematic drawing. This is done by placing code letters such as SR {slow release) in a box, drawn as part of the relay core. Illustrations of this are shown on Fig. 2-8. On the same figure is a list of the more common abbreviations used in similar fashion for indicating other special operating features of relays. Relay Contacts. The function of a contact on a general-purpose relay is to provide a reliable low-resistance circuit path for its load under s peciiied conditions of current flow, type of load, and total number

AC - allernaring

NR - nan-reaclive

currenl

D - differenrial DB - double

P - magnelically

biased

(biased

polarized, using biasing spring, or having magner bias t

in balh direclians)

SA - slow acting

DP - dashpor EP - elecrrically

SO - slaw operate

polarized

FO - fasr aperare•

SR - slow release

FR - lasr release•

SW-

MG - marginal

sandwich wound to improve to longitudinal currents

balance

TS- two step

NB - no bias

ANY COMBINATION OF THESE SYMBOLS MAY BE USED WHERE A RELAY POSSESSES MORE THAN ONE SPECIAL FEATURE. "Used where unusually

t

The proper applied

poling

lost operation

for o polarized

10 the winding

or fosl releasing relay

If lhe relay is equipped

wilh

loword

numbered

to 1he circuil

shall be shown by the use of

leads. The interpre101ion

move or lend 10 move rhe ormolure

is essenriol

+

operation.

and -

shall be 1h01 currenl in lhe direclion

lhe conlocl

terminals,

shown neoresl

1he proper

lerminol

designations

indicaled

shall

1he core on lhe schematic. numbers shall be shown.

Fig. 2-8 Symbols Used to Indicate Slow Action and Other Special Features

of Relays

The Design of Switching Circuits

20

SPHERICAL

( c,

POINT AND DISC

BAR(CROSSED)

BIFURCATED SPRING (TWIN CONTACTS)

8t?7~---

Fig. 2-9 Common Forms or Relay Contacts (Not Drawn to Scale)

of operations. To this end, contact metal in a variety of shapes and materials is fastened to the mating surfaces of the relay springs. Three common shapes of contact are shown on Fig. 2-9. Any of these may be mounted either singly or, in order to minimize troubles due to dirt, in twin form on independent or bifurcated springs. The contact metals most frequently employed are silver, platinum, palladium, and several alloys based on these and other metals. The precious metals a.re used to minimize the effects of corrosion and tarnish. When properly used in circuits and when maintained in a clean atmosphere, the contacts of well-designed relays are capable of giving good service for many millions of operations. Improper use results in accelerated erosion of the contact metal. In addition, exposed con_j tacts are always subject to the La t deposition of foreign material (film or dust) on the currentcarrying surfaces. The effect of r- -, these trouble sources is modified by mechanical considerations such CONTACT :I PROTECTION :I as contact pressure, contact chatNETWORK. 1 I ter during operation, and the wipL __ J ing action that often takes place when contacts open and close. >--... Contact troubles manifest themselves in service by high or variable contact resistance, a completely open circuit, or mechanical Fig. 2-10 Contact Protection in an Inductive Circuit locking of the contacts.

I

r

Switching Apparatus

21

Contacts are usually rated in terms of maximum steady-state current, maximum current to be made or broken, circuit voltage or power, and type of contact load. Contact erosion is accelerated by a reactive load, but the effect of such a load can be substantially reduced by the use of suitable circuit measures. An inductive load, such as a relay winding, for example, produces a high voltage by self-induction when the circuit to it is broken. This, in an unprotected circuit, causes sparking and rapid deterioration of the contact breaking the circuit. One form of contact protection consists of a resistor-capacitor network bridged across the contact or the load to absorb the high-voltage surge. A typical application is shown on Fig. 2-10. The optimum value of resistance and capacitance is a function of such variables as the inductance, circuit voltage, and load resistance. U the load is capacitative in nature, the charging current, when the contact is made, may be high enough to damage the contacts. In this case, series resistance in the contact path will reduce the initial current to safe proportions. 2.2

SPECIAL-PURPOSE

RELAYS

The function of special-purpose relays is implied directly by their name. They are the relays which are used to meet operating conditions outside the capabilities of general-purpose relays. Special-purpose relays are used primarily to respond to unusual input signals, or as the terminal equipment in signaling systems working over a distance. In some circumstances, special-purpose relays may be required as output apparatus - to handle very heavy cur-rents, for example. Except for performing certain checking functions, they are rarely used within the main body of switching circuits. Since the major portion of this volume is directed toward the logical aspects of circuit design in which special-purpose relays have little part, only a cursory review of their characteristics will be given here. It is safe to say, however, that in practice the circuit designer will find it possible to locate a choice of commercial relays to meet almost any reasonable extremes of operating conditions. Polarized Relays. Polarized relays respond to current direction in addition to current magnitude. This effect is most frequently achieved by a permanent magnet incorporated in the relay structure, although some types achieve a similar result by electrical means. Many polarized relays are characterized by high sensitivity in addition to their primary attribute.

22

The Design of Switching Circuits

The polar relay, in general, operates in one direction to close one contact or set of contacts when current of a particular polarity is connected to its winding; and operates in the opposite direction to close a different contact or set of contacts when the current flow through the winding is reversed. On sensitive polar relays, usually no more than a single transfer contact is provided. The relay may remain in its last operated position when the circuit is opened, or it may be provided with a spring bias or a separate biasing winding to restore it to a known position. Some varieties of polarized relay assume a neutral position when no current is flowing through a winding. Alternating-Current Relays. Although d-c relays will respond to alternating current, the armature has a strong tendency to chatter or buzz at twice the frequency of the applied voltage. In effect, the relay starts to release each time the current passes through zero. This difficulty can be eliminated in relays designed for a-c use by special construction of the relay core. The usual method is to split the core where it faces the armature and encircle one prong with a copper ring. The copper ring introduces a phase shift in a portion of the magnetic flux flowing through the magnetic circuit of the relay so that at no time does the total pull on the armature decrease to the point where the relay can start to release. Extremely Slow Relays. Operate and release times attainable by the use of conducting sleeves can be greatly exceeded by other special methods of construction. Two of these techniques will be mentioned here. The dashpot relay must force a volume of oil in a chamber through a small aperture before it can actuate its contacts. It can be arranged for immediate release to provide for rapid recycling. This construction permits delays with fair accuracy from a few seconds up to several minutes. The thermal relay employs a heating coil and an element which expands or shifts position under application of heat. The element, usually a bimetallic strip, actuates the contacts. Means for compensating for ambient temperature changes are normally incorporated in the structure. This type of relay can provide delays up to thirty seconds or more, but it is inherently incapable of rapid recycling. Sensitive Relays. Very sensitive relays can be built by special methods of construction, involving reduction in mass and inertia of moving parts, improved balance, light contact pressure, and close mechanical tolerances. As opposed to general-purpose relays which require operating power of the order of one to ten watts, sensitive

Switching Apparatus

23

relays are available which operate on a few milliwatts. Such relays have a minimum of contacts, but usually offer high speed in addition to sensitivity. Marginal Relays. With general-purpose relays, it is difficult to achieve a reliable non-operate current level much higher than 50% of the operate requirement value. However, by special construction, nonoperate percentages as high as 90 are possible. Mechanically Locking Relays. Relays are available with a latching arrangement which holds the armature in an operated position when the circuit to the operating winding is broken. Release is by means of an additional electromagnet, with a separate winding, which opens the latch to restore the armature. Relays of this type have little general utility, but may be of use when a circuit requires relays to remain operated for exceptionally long periods of time. Heavy Contact Load Capacity Relays. Relays sometimes must switch relatively large amounts of power as, for example, in controlling motors. For such purposes relays are manufactured with special heavy-duty contacts, or are arranged to operate mercury switches. Miscellaneous Types. In addition to the above types of relay, many other varieties with special features are available. There are vibration-resistant and shock-resistant relays for use where such conditions may be encountered in service. Midget and light-weight relays are designed for use where size or weight is an important consideration. Special construction is often necessary to achieve an exceptionally high number of operations or a very long life. For use under adverse weather or temperature conditions, relays with special finishes and coil treatments are manufactured. Although this is but a partial listing of available relay structures and types, it should give an indication of the wide variety of relays and relay-like devices available to the circuit designer. However, it is worth pointing out again that the use of the special-purpose relays is limited, and that, in the field of control-circuit design covered in this volume, the general-purpose relay is the type required almost exclusively. 2.3

ELECTRONIC

DEVICES AND ELECTROMAGNETIC

SWITCHES

As pointed out earlier in this chapter, the description of electronic tubes, semiconducting devices, and electromagnetic switches is more conveniently given in chapters dealing with their circuit usage. The discussion of tubes and semiconductors is included in Chapter 10; . the comparable description of switches is in Chapter 9.

The Design of Sw_itchingCircuits

24

:alii: :tr -t

♦

-

--v

o-LA

t

TWO MAKES SINGLE THROW NON- LOCKING

MAKE AND BREAK

A

B

t

t

LOCKING

SINGLE THROW LOCKING

DOUBLE THROW

Flg. 2-11 Convention for Push-Button or Lever-Type

2.4

NON· LOCKING

C Keys

MISCELLANEOUS SWITCHING APPARATUS

In addition to the basic apparatus described or mentioned previously in this chapter, switching circuits often require miscellaneous components to some extent. Many of these, such as resistors, capacitors, inductors, are standard electrical apparatus and require no discussion. Many other items are highly specialized types of equipment which have limited use in particular applications. There is nothing to be gained in attempting to catalogue them here. However, three types of apparatus which have frequent and important use in switching circuits are manual switches or keys, jacks, and plugs. A brief description of their functions and characteristics will be given. A switch or key is a manually operated set of contacts used to impart "on" or "off" information to a circuit, or to set up connections to or within a circuit. Although the designations "switch" and "key" are both in good standing for apparatus of this nature, the word "key" is used more frequently in discussing switching circuits.

Flg. 2-12 Typical Waler Switches

Switching Apparatus

25

Keys are available in non-locking types which restore when manual control is removed, and locking types which hold the position to which they are set. The contacts and springs of many types of key are similar to the contacts of general-purpose relays. The number of contact springs ranges from a single make or break to elaborate combinations of different contact arrangements. Typical examples of key contacts, illustrating the standard schematic conventions, are shown on Fig. 2-11. Keys are most often actuated by a lever, a push-button, a ~ ~ turn-button, or a knob. When the TWO THREE device is set by the rotary moCONDUCTOR CONDUCTOR tion of a knob, the designation A B "switch" is used more frequently than "key". The number of sepaFig. 2-13 Plug Conventions rate key positions, each with its own contact set, depends upon the type of key or switch. Push-button keys usually have two positions; lever-operated keys, two or three positions, as shown on Fig. 2-11; and rotary switches may have up to twelve or more. The latter switches operate on a different principle from lever or push-button keys, usually rotating brush or wiper contacts over wafers bearing fixed contacts or contact strips. Examples are shown schematically on Fig. 2-12. Jacks and plugs are used for setting up temporary connections on a flexible basis. For example, a number of circuit paths may be con.:. nected to a field of jacks, and any desired interconnection between these circuit paths may be established manually by, inserting cordconnected plugs in the corresponding jacks. Standard pl u g s, which carry two or three conductors, are shown schematically on Fig. 2-13. The corresponding jacks are sh own on Figs. 2-14A and 2-14B. In many circuits utilizing this type of connection, it is des i r able to have the insertion of a plug open or close circuit paths that are independent of the paths through the plug. For this purpose, auxiliary springs may be supplied on the jacks, as shown on Fig. 2-14C.

~

~

TWO CONDUCTOR

CONDUCTOR

A

B

THREE

7

~ ~

JACKS WITH AUXILIARY

SPRINGS

C Fig. 2-14

Jack Conventions

The Design of Switching Circuits

26

1

I BATTERY SYMBOL

GROUND SYMBOL

GROUNDED BATTERY

A

B

C

__l 1

-=--e· J;.,._...,!oNTROL T...._.... CONTACT

CONTROL CIRCUIT FOR A RELAY

EQUIVALENT CONTROL CIRCUIT FOR A RELAY

D

E

Fig. 2-15 Schematic Conventions for Battery and Ground

2.5

BATTERY AND GROUND CONVENTIONS ON CIRCUIT DRAWINGS

Before leaving the subject of switching apparatus, the battery and ground conventions that are in general use on circuit drawings and that will be used in this book should be explained. Several of those conventions are illustrated by Fig. 2-15. On Fig. 2-15A is shown the standard battery symbol with polarity indicated. The battery voltage may be shown where necessary. In text illustrations, the positive and negative battery terminals will be indicated solely by the long and short lines, respectively, of the symbol. It is customary in most circuit applications to connect one side of the battery supply to the earth, or ground, in order to insure that the system is tied to a fixed potential. The symbol for ground is shown in Fig. 2-15B, and the symbol for a grounded battery (positive side grounded) in Fig. 2 -15C. Usirig this symbolism, the complete operating circuit for a relay may be shown as in Fig. 2-15D. However, if it is understood that a common buss connection makes available the same ground potential at all parts of the circuit, the drawing may be simplified as shown in Fig. 2-15E. Because of the resultant reduction in the number of lines on the circuit drawing, this convention of separated battery and ground will be used throughout the text.

Chapter 3 BASIC RELAY CONTROL PATHS

A relay circuit is an interrelated set of relay structures with input and output paths and internal configurations of contact networks. The relays are controlled both by the input paths and the contact net­ works on the relays themselves, whereas the output paths are activated for the most part only by the contact networks. Additional elements such as resistors, capacitors or varistors may occasionally be intro­ duced into the relay circuit to achieve particular results. Although the term "circuit" is sufficiently general to include groups of relays ranging in number from one to several thousand, the relays of a circuit usually are related functionally. If larger circuits are well designed, they break down naturally into smaller functional units each of which has all the attributes of a complete circuit, that is, inputs, outputs, and contact networks. A further breakdown of a circuit yields the actual control paths which operate and release each of the relays at the proper time. Although these control paths may be complex networks involving perhaps dozens of individual contacts, they have the common characteristic that they may occupy but one of two states at a time, that is, closed or open. The condition where an impedance is inserted in a control path is considered a special case which will be treated separately. The methods of designing contact networks to meet circuit re­ quirements are taken up in the next two chapters. Before the implica­ tions of contact network design can be adequately comprehended, however, it is necessary to understand the various ways in which a relay can be operated and released by an open or closed path. This chapter discusses these fundamental factors in relay circuit design.

3 .1

BASIC CONTROL METHODS FOR RELAYS WITH SINGLE WINDINGS

There are four basic methods for controlling a relay with a single winding. Two of these, each the inverse of the other in effect, permit operation and release of the relay by closing or opening of a path. The other two cannot operate the relay by themselves, but permit holding it in an operated or released state by closing a path. - 27 -

The Design of Switching Circuits

28

A

B

Plg. 3-1 Direct Control Operation

Direct Control. Fig. 3-lA shows the simplest and most common relay control path. Relay (X) is under direct control of contact network A: when A is closed, (X) is energized and operates; when A is open, (X) is de-energized and releases. Although Fig. 3-lA and Fig. 3-1B are the same in effect, the arrangement in Fig. 3 -lA is generally preferred in practice since false grounds on the contact network will not blow a fuse. The arrangement shown in Fig. 3-1B is sometimes desirable, however, when other requirements of a circuit are considered. It is important to remember that in this and all other relay circuits there is a time delay between path closure and relay operation while the magnetic field is being built up and the moving parts of the relay are in transit. A similar but not necessarily equal delay exists between the opening of the path and the release of the relay.

J

[P

(X)

'+-+---1 ........ ~---1

I r---:i_ I

A

-1-Flg. 3-2 Shunt Operation

Shunt Control. In some cases it may be desired tooperate a relay when a path opens. The circuit of Fig. 3-2 does this by using a shunt path. When A is open, (X) is energized; when A is closed, (X) is deenergized. Although this arrangement is sometimes very useful, it has several disadvantages. In relay circuit design, much effort is spent to keep a circuit economical and reliable. The resistor shown in Fig. 3-2 is necessary to limit current through the shunting path, but adds to the

Basic Relay Control

29

Paths

expense of the circuit. The current drain through the shunt path performs no useful work. In addition, the release time of the relay is increased since current in the winding decays exponentially through the shunt path, instead of terminating abruptly.

0

o B

Fig. 3-3

♦

o-----tb ::- l

r

+i

1

Flg. 3-4 Variation of Basic Lock-up

Lock-up

Lock-Up. In the previous control methods, the controlled relay had no part in its own operation. In a "locking" circuit, a relay affects its own control path. Thus, in Fig. 3-3, the path which operates relay (X) is through network A. Thereafter the relay can hold through network B, although prior to the operation of (X), network B has no effect upon the circuit. Relay (X) is said to have locked-up to B through its own contact. It is this circuit that imparts to relays the quality of "memory." That is, the relay is enabled to hold operated after the operating condition has disappeared. Fig. 3-4 shows a similar locking circuit where A and B control the relay in a slightly different manner. TO OTHER CIRCUITS

°1-

oo

D

( X)

11~

--+-+-

Fig. 3 -5 Continuity Lock-up

Another method of lock-up using a continuity spring combination is shown in Fig. 3-5. Once (X) is locked up under control of B, A is free to control other paths without being affected by the ground on the winding of (X). It is evident that in this arrangement B must always be closed before A, or the act of (X) operating will break its own operating path and release itself. Relay (X) will then buzz• until Bis closed, locking it up. Other combinations of lock-up and operate paths which have different effects in the circuit are shown on Fig. 3-6. • If the back-contact of a relay ls Included In Its own operating path and no early-closing locking path ls provided, the relay will alternately operate and release at a rapid rate when current ls applied. This ls known as buzzing.

The Design of Switching Circuits

30

-r

TO OTHER 8 0-

...,-0

A o-----CIRCUITS

Flg. 3-6 Varlations of Contlnulty Lock-up

Lock-Down. Lock-down is a scheme whereby a relay enters into its own control to keep itself from being operated by one path until another path has opened. As seen in the two arrangements in Fig. 3-7, network A cannot operate relay (X) unless network B is open. Once (X) has been operated by A, however, B may open and close with no effect resistor, and it will upon (X). This circuit requires a current-limiting drain current while the relay is shunted down. The acting time of the relay may be affected slightly by addition of the current-limiting resistor; the shunt lock-down path, however, does not change this acting time.

B

o---:t_

-=Flg. 3-7 Lock-down

3.2

SUMMARY OF THE FOUR BASIC CONTROL METHODS

Fig. 3-8 shows the four basic control paths applied to one relay. The paths are the direct D, shunt SH, lock-up LU, and lock-down LD. Analysis will show the interrelationships between paths. For example, the relay can never operate if D is open, or SH is closed, or LD is closed; SH can always release the relay; LD is without control after the relay has operated, and so forth. In circuit design it is often necessary ~ LU ~

SH7-

Flg. 3-8 Comblned Basic Control Paths

Basic Relay Control Paths

31

to interchange control paths by contact network manipulation in order to achieve the desired operating characteristics. 3.3

RELAYS WITH SPECIAL CHARACTERISTICS

The preceding paragraphs have dealt with the basic ways of controlling a single-winding, general-purpose relay. Often,situations arise within a switching or control circuit that can be met more economically by relays with special characteristics than by relays controlled only by the paths described above. In addition, a restrictive design condition, such as a limited number of leads available for control, can sometimes be met only by use of relays with special characteristics. The rest of this chapter deals with the control of such relays. -----o----o

TO OTHER CIRCUITS

°]_ -=-

( X)

A

a

B

Fig. 3-9 Use of Multlwound Relays

Relays with More Than One Winding. The most common use of a relay with several operating windings is to permit two or more operating paths of this relay to be isolated. A simple case is shown in Fig. 3-9A. Either or both paths will operate relay (X) without shunting or otherwise interfering with each other. This scheme can be used to isolate completely an operating path from a locking path without the strict time sequence required of a continuity lock-up, as in Fig. 3-9B. A useful arrangement employing a double-winding relay is the differential circuit. In this circuit, the two windings are connected in oppcsition so that when both are energized, the net flux through the core is insufficient to operate or hold the relay. As shown in Fig. 3-10, A or B (X) alone can operate relay (X); but if neither or both are closed, relay (X) is released. With this circuit, the two Windings must be closely equal in turns and resistance, or external resistance must be added to balance the effect of the windings. A common Fig. 3-10 DWerentlal Relay

The Design of Switching Circuits

32

(OPR)

(OPRl A

_c---o

_C"°

-

I

A 0--��--h

IRLSl

_r--11 1--o 8

B

A Flg. 3-11 Forced Release or a Relay

use for this arrangement is to indicate if two circuit paths such as A and B do not open and close at the same time. Another use is to permit releasing a relay by closing instead of opening a path. The operating path must remain closed until the release path is closed, and then the two paths must be opened simultaneously. A variation of this scheme is to use an opposing release winding which is more effective than the operate winding in developing a mag­ netic field. By this means the total flux in the relay core can be driven rapidly through zero, producing a fast release action. In order to insure that the current paths are opened before the relay can re-operate on the reversed field, a make contact on the relay in question should be included in the circuit as shown on Fig. 3-11.

(MG)

Flg. 3-12 Operation

or Marginal and Sensltlve Relays

Marginal Relays. A marginal relay is one that will not operate on currents below a definite value through its winding, but will operate on currents over this value. It is often used in arrangements such as Fig. 3-12 where, in combination with a sensitive series relay, two condi­ tions over. a single lead can be detected. The sensitive relay operates on current below the non-operate current of the marginal relay. Thus A alone closed operates (S), but resistance R keeps the current too low to operate (MG). B closed operates both (S) and (MG). This circuit is one of the very few in which appear control paths with more than two conditions. Theoretically a number of marginal relays could be used in series, operating successively as the current is increased in steps. However, variations in apparatus, adjustments, heating, battery voltage,

Basic Relay Control

Paths

33

and so forth, make close· margins hard to meet; practical circuits are seldom found having more than one marginal condition. Polar Relays. Polar relays have two characteristics which make them valuable in switching and control circuits. First, they react to current direction; second, they are Fig. 3-13 quite sensitive and capable of operPolar Relay with Blasing Winding ating at high speed. The spring combination of many polar relays is limited to a maximum of a single transfer, and, unless biasing springs or windings are provided, there is no assurance which side, if either, of the transfer will be closed when the winding is de-energized. Because of their sensitivity, polar relays are often used as marginal or sensitive relays to meet critical circuit conditions. An arrangement is shown in Fig. 3-13 where the relay has two windings, one of which is used for biasing. The biasing winding is poled to oppose the operating winding, and when energized alone, it holds the armature against contact No. 1. When the current in the operating winding exceeds that of the biasing winding by a small amount, the armature will operate to contact No. 3.

I t

0

L +fi

1

/ --~

+

-

I

r~/ A.

lib_ -

B

A

Fig. 3 -14 Polar Relay Control

When the relay is used for its polar characteristics, some means of reversing the flux must be provided. Fig. 3-14A shows an easy way of doing this if two battery supplies are available. Two voltages of different amplitude but the same polarity achieve the same result, as shown in Fig.3-14B. The source of lower voltage might be a resistance voltage divider. Slow-Acting Relays. Slow-acting relays include a deliberate factor of time added to the operation or release of a relay. The slow action may be desired for either of the following two reasons:

The Design of Switching Circuits

34

1. To produce or measure a definite time interval. 2. To insure correct sequence of operation in a multirelay circuit. Slow-acting relays are, in general, controlled in exactly the same manner as standard relays, and can be incorporated into relay circuits with no difficulty. When a slow-release relay is used for timing, it is necessary to pre-operate the relay sufficiently in advance of the timing period to insure that current through the winding can attain full value before the relay control path is opened.

PROBLEMS FOR CHAPTER 3 3-1

A relay (R), equipped with only a single transfer-contact, operates when a key (A) ls operated and remains operated when the key ls restored to normal. The relay may then be released by operating a key (B), and will remain released when B is restored to normal until (A) ls agaln operated. Show two circuits for each of the four following conditions where the springs or the keys are llmlted to a single make or break on each key. For one circuit, one contact spring of both keys must be permanently connected to ground; for the other clrcuit, a contact spring or one key only may be permanently connected to ground. Use only slngle-wlnding relays.

3-2

(a) Key (A): BREAK Key (B): BREAK

(c) Key (A): Key (B):

{b) Key (A): MAKE Key (B): MAKE

(d) Key (A): MAKE Key (B): BREAK

Three relays (X), (Y), and (Z) are controlled by keys (A), (B), (C), and (D) as follows: relay (X) operates when key (A) ls operated and releases when key (A) ls released; relay (Y) operates when elther key (A) or key (B) ls operated and remains operated untll key (D) ls operated; relay (Z) operates when any or keys (A), (B), and (C) are operated, and remalns operated until key (D) is operated. Key (D) ls only operated when all other keys are released. Keys (A), (B), and (C) are each provided wlth a single make-contact and key (D), with a slngle break-contact. Design a circuit fulfilling contact per relay.

3-3

BREAK MAKE

these requirements

and using no more than one make

(a) Two non-locking keys, (A) and (B), each with two make-contacts, are operated at random. Design a circuit for operating a ·relay (X) when either key is operated, but not when both are operated. The relay ls to remain operated untll both keys are operated; at which time the relay is released. (b) Repeat part (a), assuming that each key ls equipped with two break-contacts only.

3-4

In Flg. 3-10, assume that the contacts A and B controlling relay (X) are make contacts on relays (A) and (B). Design a circuit for controlling a single-winding relay (X) by contacts on (A) and (B), so that (X) operates in the same maMer as in Fig. 3-10. Additional contacts may be used on relays (A) and (B) if necessary.

35

Basic Relay Control Paths 3-5

Two relays are selected of 0.3 second.

to have an operate time of 0.2 second and a release

(a) Design a continuously ruMing interrupter circuit which will give put pulses grounded for 0.6 second and open for 0.4 second.

time

repetitive out-

(b) Arrange the above circuit so that operation of a control key will change the output pulses to ground for 0.5 second, open for 0.5 second. 3-6

Three keys, (A), (B), and (C), control a relay (X). Relay (X) should operate when any one key is operated or when all three keys are operated. Under all other conditions, the relay should be released. Design the circuit, using no more than one transfer contact on each key.

3-7

In the diagram given below, indicate the positions in which lamps a, b, c, and d light as switch (SW) is turned from 1 to 8. Relay (B) will not operate or remain operated in series with relay (A) and a resistor (R) on 24 volts. Indicate also what relays are operated for each position of (SW). (R)

2 3

-~I\ (SW)

)4

-::(Bl

5

(Rl 6

7

I48VOLTS

-=-

8

(Al

Chapter 4 RELAY CONTACT NETWORK CONFIGURATIONS

A relay contact network is an arrangement of a number of individual contacts jointly controlling circuit paths. It is convenient to divide contact networks into two general types: two-terminal networks with a single input and a single output terminal; and multi-terminal networks, usually with a single input and a plurality of outputs. The sequence of operation of the individual contacts in the network may or may not be of importance, depending upon the application of the network. In this chapter, problems concerned with the configuration of contact networks are considered, without regard to sequence of operation. The network manipulations and simplifications discussed are of the most elementary variety. The problem of systematic network rearrangement is taken up in much greater detail in Chapter 5.

4.1

LOGIC OF CONTACT NETWORKS

The requirements which a contact network must fulfill may be formulated by stating those conditions under which the network will be open and those under which it will be closed. An analysis of the state ment from the standpoint of logical validity is equivalent to an investigation of the network configuration itself. Thus, the tools and methods of logic underlie the analysis and solution of network problems. Any two-terminal contact network can be considered to be a twovalued device. At any given instant it presents either an open (infinite impedance) or closed (zero impedance) path to electric current. Thus, the network can be represented by a single contact which is open or closed under conditions identical to those imposed on the network. Because the network, like the contact, can only be either closed or open, it is similar to a proposition in logic which is either true or false depending upon the truth or falsity of the various conditions of the proposition. The closed condition of a contact or network can be considered the equivalent of a true condition in logic. Then a proposition which is true only if conditions A and B are both true corresponds to a network which is closed only if contacts A and B are both closed. Each condition upon which the proposition relies corresponds to an individual contact in the network. The design of a contact network to - 36 -

Relay Contact Network Configurations

37

perform a certain function under given conditions corresponds to the problem in logic of synthesizing a complete proposition from individual true or false statements. Moreover, the arrangement of the individual contacts in the network is similar to the relation of the various statements whose truth or falsity make the proposition true. If the synthesis of the proposition is not completely sound, the corresponding contact network will not function in the required manner. This similarity between contact networks and propositions in logic will be more fully brought out in succeeding sections.

_r

A

00----00

B

Fig. 4-1 Relay (R) Operates for A and B Closed

4.2

0--

0-

Fig. 4-2 Relay (R) Operates for A or B Closed

"AND-OR" AND SERIES-PARALLEL

RELATIONSHIPS

The most basic of all two-terminal networks are the simple series circuits and simple parallel circuits shown in Figs. 4-1 and 4-2. The network of Fig. 4-1 is closed, and relay (R) operates, only when both A and B are closed; whereas in Fig. 4-2 the circuit is closed if A or B (or both) are closed. Fig. 4-1 represents a proposition which is true only when both conditions A and B are true. If only one of the two conditions is true, the proposition is not true, and the circuit in Fig. 4-1 would not be closed. A proposition which is true if condition A or B is true is represented by Fig. 4-2. It is evident that the conjunction "and" is equivalent to a series contact arrangement, and that the conjunction "or" is equivalent to a parallel contact arrangement. For example, consider the circuit shown in Fig. 4-3. Relay (R) operates if A and (B or C) are closed. The equivalent statement in logic would be: "The proposition is true if conditions A and (B or C) are true." Similarly, in Fig. 4-4 the circuit is closed (or the proposition is true) if A or (B and C) are closed (true). Although for more intricate circuits the punctuation of the statement might become highly complicated, it is possible to describe

r 7: : A

0---

n1,1-

Fig. 4-3 Relay (R) Operates for A Closed, and B or C Closed

F •·."J

0--

,.,Y.1,1

Fig. 4-4 Relay (R) Operates for A Closed, or B and C Closed

The Design of Switching Circuits

38

0-----

_[ oo· (Al

_I_

,.,-0-i+-i

L-.ln:J

0

- 1e,O

1c,O

Fig. 4-5 Relay (R) Operates for (A) Operated, or (B) and (C) Operated

completely the conditions for circuit closure for any series-parallel network by indicating parallel combinations by "or" and series combinations by "and". It should be noted that in the phrase "A or B", the two components are not mutually exclusive. The network-in Fig. 4-2 -is closed if A or B or both are closed. In dealing with contact networks, the term "or" should always be interpreted to imply "either or both" unless otherwise explicitly stated. Rather than stating a set of requirements in terms of open and closed contacts, it may be more convenient to state it in terms of relays operated and released. If a circuit is to be closed when a relay is operated, a make-contact is indicated; if the closure results from the release of a relay, the condition requires a break-contact. The relay circuit in Fig. 4-5 is equivalent to the network in Fig. 4-4 since the operating path for relay (R) is closed when relay (A), or relays (B) ·and (C), are operated. Another representation is shown in Fig. 4 -6, where relay (R) operates when relay (A) is operated, or when (B) is released and (C) is operated.

l

D

-=- (Alo

J

·-------------

_L

0

,---L

0

(Bl

(Cl

IRJ@-1,ri -=-

D

Fig. 4-6 Relay (R) Operates for (A) Operated, or (B) Released and (C) Operated

4.3

THE NEGATIVE RELATION

In logic the "negative" of a proposition is a proposition which is false for those conditions, and only those conditions, for which the original proposition is true. The original proposition and its negative are therefore mutually exclusive. Similarly, the relation between a contact network and its negative is that one is closed when the other is open. It follows that, if network A is the negative of network B, B is

39

Relay Contact Network Configurations

also the negative of A. The most elementary example of a contact network and its negative is a make-contact and a break-contact on the same relay. When a circuit through one contact is closed the other is always open•. It is often possible to determine the negative of a particular network by inspecting the statement of conditions for closure for that network. A

A

A 0-1_-o-----i-~---~

B

B Fig. 4-7 and Its Negative

A Simple Network

B

B

C

Fig. 4-8 Another Network and Its Negative

The circuit in Fig. 4-7A is closed only when relays (A) and (B) are both operated. The negative of this network would then be one which is open only when relays (A) and (B) are both operated, and closed when (A) or (B) is released. In this final phrase, the "or" suggests parallel contacts, and the fact that the relays are released to close a circuit suggests that the contacts be breaks. The circuit is shown in Fig. 4-7B. A second example is that of Fig. 4-8A where relay (A) must be released and either (B) or (C) operated to close the network path. Brief consideration indicates that the circuit in Fig. 4-8B is the desired negative. From examination of the above networks and their negatives, it appears that to replace a network by its negative, it is only necessary to substitute for each contact in the network its negative and to place these negatives in a network in which series and parallel connections have been interchanged. In actuality, this method of obtaining negatives holds for any two-terminal network consisting of only series and parA somewhat more complicated network and its allel connectionst. negative, illustrating the method, are shown in Fig. 4-9. • This, of course, ls not strictly true with commercial relays. During these relays one contact may close before or after another opens. tThe proof justifying

this method will be indicated

in Chapter 5.

the acting tlme of

The Design of Switching Circuits

40

D

A

E

C

A

I

~

;:

t

D

B Fig. 4-9 A More Elaborate Example

4.4

ora Network and Its

Negative

BASIC SIMPLIFICATIONS IN TWO-TERMINAL NETWORK DESIGN

Designing a network to meet requirements by substituting a contact for each logical condition often results in a redundance of contacts. After the two-terminal contact network has been designed for closure under given combinations of relays operated and released, it may be possible to simplify the network by combining certain contacts and, perhaps, omitting others. First inspection often indicates simplifications which may be made at once. In the network of Fig. 4-lOA, the circuit is closed if relays (A) and (C) are operated, or if relay (A) is operated and (D) is released. It can easily be seen that an equivalent

A

A 0

C

A

A

B D

B Flg. 4-10 Combining Contacts

•..----..---00

L_fct...__l B

Fig. 4-11 A Second Example of Combining

Contacts

41

Relay Contact Network Configurations

but simpler statement of the same conditions would be: relay (A) operated, and (C) operated or (D) released. The network in Fig. 4-l0B, then, is equivalent to that of Fig. 4-l0A. A similar example appears in Fig. 4-11. Here it is again possible to combine the two make-contacts on relay (A) into a single make.

A

-1_0 0

A

t

B

B

0

C Fig. 4-12

Network Simplification

Illustrating

an Invalid Step

When simplification of the type illustrated above is done, it is well to check that the final network contains all those paths and only those paths that appear in the original network. Fig. 4 -12A shows a network, and Fig. 4-12B, a simplified equivalent to that network. An attempt to simplify further the network of Fig. 4-12B might suggest combining the two makes on relay (C), resulting in the circuit of Fig. 4-12C. This last manipulation, however, cannot be justified since a path not inherent in the original circuit is present in Fig. 4-12C. (The network of Fig. 4-12C is closed when B alone is closed, which is not true of Fig. 4-12A). Other basic simplifications become evident when the actions of networks are closely analysed. For example, requirements might state that a network should be closed when contacts A or (A and B) are

The Design of Switching Circuits

42

closed. If the network is drawn and examined, it can be seen that the network always is closed when A is closed, no matter what the condition of B is. Therefore the expression "or (A and B)" is redundant and the network reduces to a single make contact, A. A similar result is obtained by analysis of the network to be closed when contacts A and (A or B) are closed. Another example of permissible simplification occurs when a contact of a network is in series or parallel with a portion of the network including the negative of the first contact. In this class is the

E A

8

C

C

A

A A

C

A

8

D

B

'

0

0

A

I ;: J 8

0

D

C

0

f

0

E

I

2

·~d D

Fig. 4-13 Simpllflcatlon Employing Negatlve of Network

0

Relay Contact Network Configurations

43

network closed when relay (A) is operated or (relay (A) is released and relay (B) is operated). This network is always closed when (A) is operated, and, when (A) is released, can only be closed by operation of (B). Therefore the break-contact on (A) representing (A) released is redundant, and the network reduces to A or B. This type of simplification will be discussed more fully in Chapter 5. An indirect method of simplification, which in some cases may indicate steps not otherwise obvious, involves deriving the negative of the network, simplifying the negative, and finally taking the negative of this simplified circuit. In accordance with previously defined relationships, the negative of the negative of a network is exactly equivalent to that network. To illustrate, the negative of the network in Fig. 4-13A is determined by inspection. This negative circuit, shown in Fig. 4-13B, is simplified by inspection to give the network of Fig. 4-13C. Observe that the uppermost branch in Fig. 4-13B is unnecessary since it is paralleled by a branch of series breaks on relays (A) and (C) which is unaffected by the condition of (B). Then, the negative of this simplification is drawn as in Fig. 4-130. Examination of Fig. 4-130 shows that it contains all those paths contained in the circuit of Fig. 4-13A and no others. A o-o ---oo

A o-o

E o--o

C o-o ---oo C o-o

A

Fig. 4-14

4 .5

B

D D

E o--oB

B

Basic Bridge Network and Its Serles-Parallel

Equivalent

BRIDGE -TYPE NETWORKS

Not all two-terminal networks are composed of only series and parallel elements. Sometimes a circuit arrangement will include crossconnecting paths between parallel paths of series elements. Such networks are referred to as bridge-type networks. The fundamental bridge-type network, shown in Fig. 4-14A, is conveniently analyzed by tracing all possible paths between input and output terminals. These paths indicate all the conditions for closure for the network and may be sketched as in Fig. 4-14B. Any bridge-type network may be transformed into a series -parallel arrangement by this

The Design of Switching Circuits

44

-

IA

0

I

71

OD

C

J

v:::

Leo

F

O

OGL

E

l

oHr

Fig. 4-15 A More Complicated Brldge-Type Network

process of investigating all paths. The number of paths increases with the complexity of the bridge design. For example, the bridge in Fig. 4-15 has thirteen possible paths between input and output terminals. It is evident that the bridge circuit may afford some economy in number of contacts. Thus, such a network may be superior in this respect to its series-parallel equivalent. However, very few seriesparallel arrangements have an equivalent bridge and, moreover, the equivalent bridge may be very difficult to derive. Contact network simplification may, if the designer is not on guard, lead to a bridge-type circuit which is not exactly equivalent to the original network. An illustration of such inappropriate simplification is· shown in Fig. 4-16. Although the bridge network of this figure contains all the paths present in the original network, it also is closed a for the combination of relay (B) and (C) operated and (E) released, path not existent in the unsimplified circuit. Of course, the bridge network might be acceptable if such a condition never exists.

A

A

i_____Je7 ____

.....

♦

--

C

i______fo7

______ ...

A

E C

Fig. 4-16

Serles-Parallel

Network and Non-Equivalent

Brldge

Relay Contact Network

l

0

A

Configurations

45

L___fe1

__r--L___jci

(X)@

,,,~

(Y)o

,1,7

A

l

0

B

t

A

H-l

(X)o

C

t (Yl

0

,1,~

B

• Fig. 4-17

4.6

Three-Terminal

MULTI-TERMINAL

Network

Reduced from Two Two-Terminal

Networks

NETWORKS

The most common form of multi-terminal network is that of a single input which may be connected through the network to two or more outputs. The design of such a network may be carried out as the design of several inter-related two-terminal networks. An elementary case is illustrated by Fig. 4-17 where two make-contacts on relay (A) are combined into a single make. The resulting circuit is a three-terminal network, one terminal connected to ground and the other two to relay windings. However. care must be exercised in combining contacts to avoid "sneak" paths between outputs. Such a sneak condition is introduced in Fig. 4-18, and consists of path B and C and D in Fig. 4-18B which does not exist in the unsimplified network of Fig. 4-18A. It should be noted in the example in Fig. 4-17B that if relays (B) and (C} are operated, the windings of relays (X) and (Y) are tied together, a situation which cannot occur in the circuit of Fig. 4-17A. Whether this situation is undesirable depends upon the requirements for the circuit and possible conditions at the network terminals. For example, the addition of a locking contact on relay (X), as in Fig. 4-19, would permit a sneak path through the make-contacts on relays (C) and (B) to hold relay (Y) operated alter relay (A) has released. The effect of the sneak path is to add a second condition for the operation of relay (Y): relays (B) and (C) and (X) operated; whereas the original operating path for (Y) was through operated relays (A) and (C).

The Design of switching

46

l

0

A

L____F7 (Xl

l

0

A

L___fc7 _[

0

D

t

I

(Y)-0

0

Circuits

N-7_ ,1,r-i_ -

A

l

l

0

A

♦

_[ "T

D

D

t

~

t (Xlo

C

,1,r-i -

.J

(Y)u

,1,r-i -

B Fig. 4-18 Reduction or Two-Terminal Networks with a Sneak Path Introduced

4.7

TRANSFER CONT ACTS IN NETWORKS If two paths of a network pass through contacts on the same relay, one through a make-contact and the second through a break-contact, theoretically these paths can never be closed at the same time - as suining no contact stagger. Paths exhibiting this characteristic are said to be "disjunctive." By network manipulation it may be possible to arrange the make and break contacts so that each has a terminal connected to a common junction point; under this condition a transfer contact may be substituted for the make and break. Such a substitution is shown in Fig. 4-20. While a make-and-break-contact on a relay and the corresponding transfer combinations are both inherently disjunctive devices, the transfer combination is often preferable since it employs one less spring than the make and break arrangement. Although contact stagger may allow independent make- and break-contacts to be closed simultaneously for a short period, the transfer permits only one of two paths to be closed at a given time. Thus, it can be used to insure disjunctivity between paths or outputs.* • The designation "transfer" as used herein implies the definite break -before - make sequence. Some types of general-purpose relays employ the physical form of the transfer combination, and yet permit an indeterminate sequence. Use of such combinations must be avoided when simultaneous closure or make- and break-contacts can introduce hazards.

Relay Contact Network

47

Configurations

B

Fig. 4-19

C

A

-

A Possible "Back-up"

: l

Situation

~

1111

l

c[}

1111

l

(X)o

,1,1

l

(Yl@

1111

l

(X)

0

B

J

0

A

t

0

B

t ( Yl

A

C

A

-

: I 0

B

t

B Flg. 4 -20 The Transfer

Contact Used to Isolate Outputs

The Design of Switching Circuits

48

~

A

A

C

A

I' J

~

l

0

B

Fig. 4-21 Transfer Contact Used to Replace a Make and a Break Contact

Often it is not immediately recognizable that a transfer combination can be employed. An example is shown in Fig. 4-21 where the make- and break-contacts on relay (A) can be replaced by a transfer. Here, the transfer arrangement is probably preferable since definite sequence between make and break on (A) is established and, as a result, there can be no momentary undesirable path in the circuit due to con tact stagger on relay (A), if the make closes before the break opens. A more elaborate example of network manipulation is shown in Fig. 4-22. It is evident from this illustration that the disjunctive character of the transfer allows simplifications which, at first glance, seem invalid.

A transfer contact may sometimes be substituted for a makecontact or a break-contact in a network to permit simplification in another part of the network. The advantage gained from this substitution arises from the disjunctive aspect of the transfer. For example, in Fig. 4-18 a sneak path through contacts D, C, and B to the winding of (X) was introduced when the two A contacts were combined. In the original network, the path to (Y) is closed when contacts A and C, or contact D, are closed. Since the path through A and C is important only when relay (D) is released, a break contact on (D) may be inserted in series with A and C without disturbing the circuit. This permits arranging the circuit as shown in Fig. 4 -23, where the sneak path no longer exists. This simplification technique is applicable when it is desired to combine in whole or part one branch of paralleled groups of contacts with another part of the network. Care must be exercised in choosing the contact to be converted to a transfer, or the network may be com pletely changed in performance. The effect of the path discontinuity

Relay Contact Network

Configurations

49

introduced during the transit time of the transfer must also be considered. Methods for performing this type of network manipulation will be • discussed more fully in the next chapter.

-.oA-~-----------

,.__A_L_jat_Jo ~-___. H

~t_____Ft

__

__.

E

A C

E

H

t

♦

.G

F

B Fig. 4 -22 More Elaborate

Network Slmpllflcation

a

-f'. • A •~-_,.j

: C f

(Xl

{}--1,~

~I•~ Fig. 4-23

Use or Transfer

to Eliminate

Sneak Path

or Fig.

4-18

50 4 .8

The Design of Switching Circuits TRANSFER TREE CIRCUITS

A "transfer tree" is a particular type of multi-terminal network in which the single input may be connected to any one of m outputs according to the combination in which the relays actuating the network contacts are operated. The circuit of a four-relay tree is drawn in Fig. 4-24. There are sixteen outputs, each output corresponding to one of the sixteen possible combinations in which four relays may be operated, including the all-released condition. One output only can be connected to the input at a time, and, because of the transfer-contact combinations, the outputs are completely isolated under any condition. As well as being disjunctive, each path from input to an output in a true transfer __

I I

I I

.,., NONE

~o

~c

I

I

~co I __ _,B

I

I •

I '----o

..__oO-L...L

BO

l~Bc

♦,_____,...o

I

B co

I

I

--~A

I

L-..o

-->--+-I.......

It

: i

l

I

I I I 11il

---o-~ O

□•

o

1 1

I +~-~o

t_~_♦

~B

~C

-

Fig. 4-24 The Baslc Four-Relay

AC'

L-..o ACO _ _,o AB I .t 1

I

I

--.no ,__-.co._,,_ ....... ♦

A0

1--L--'°o

o :

~~ABCO

Transfer

Tree

ABO AB c

51

Relay Contact Network Configurations tree must pass through a contact on each controlling relay once and only once. Because of the triangular sym metry of a fully developed tree, there is a definite relationship between the number of controlling relays, the number of outputs, the maximum number of transfers on one relay, and the total number of transfers in the tree. The table at the right indicates this relationship.

Number of relays

In tree = n

Possible

re 1a y combinations

= output

terminals

Maximum

= m = 2n

number of transfers

on one relay = m/2 = 2B

I

O I -

-c0-.: _..J

! L..o BO

~BC '--'eco A

Aeo-J AC~ Aeco-J: ADC-,: I

ABoo---.J

:

o

°7J.___J

ACD

Aecoo-J

II

'

0

I

I

~A

Fig. 4-25 Rearrangement

~B

~c

~D

of Transfer Tree to Equalize Contacts

reconnected to the rest of the tree. Fig. 4-26C shows that relay (D) now has seven transfers and relay (C), five. The contact manipulation of the circuit of Fig. 4-26C may be continued as desired. In Fig. 4-27, the upper branch on relays (B), (C), and (D) of Fig. 4-26C has been rearranged by switching the transfers on (D) and (B). Note that the process of tree rearrangement is merely one of interchanging designations of the contacts composing a minor tree; in Fig. 4-27, designations B and D have been interchanged. If this rearrangement is replaced· in the circuit of Fig. 4-26C, the load distribution is: relay (A), one transfer; (B), four transfers; (C), five transfers; and (D), five transfers. Thus, by continued rearrangement of branches, the load may be shifted as desired, remembering that the total number of transfers will remain 2n -1. Similar manipulations may be performed on incomplete trees. With these circuits, it is sometimes possible to reduce the total number of contacts by rearrangement. A three-relay tree with five outputs is shown in Fig. 4-28 together with a rearrangement of the tree. Here, the designations of relays (B) and (C) are interchanged and, as a result, the contact configuration has been simplified. Care must be taken to insure that the tree obtained by manipulation has the same outputs as the original tree. Also it should be kept in mind that, in an incomplete as

Relay Contact Network Configurations

53

well as a complete transfer tree, each path from input to an output must pass through a contact on each controlling relay once and only once. _____.r--NONE

----o----1 B

r---------1 C

♦

• lt--o

t

a D

A

I

I re

~

co

---.-r--NONE

______

_i

. Le,

C

t

0

!1

f

0

a

I

c

B

L-co

--------ONONE oc D

oco

It

t

0

I t-I I

I

I

I

I

t

I

L-oeco A

o I

I L._,ACO AB o I l

I

I

0

I L-oAO lAC

I I I

-o-+

B

I L-oeo I 1 ec I

I

I

C

I I

t

Flg. 4-26 Manipulation of Tree Contacts

L.....oAeo ABC

D

The Design of Switching

54

Circuits

,----------8

three relays

(X), (Y), and (Z) as

Relays (A) or (C) in operating normally operate relay (Y). Should relay (B) be operated, relay (A) in operating operates relay (X) instead of relay (Y), while relay (C) in operating still operates relay (Y); Should relay (D) be operated, relay (C) ln operating operates relay (Z) Instead of relay (Y). With (D) operated, relay (A) in operating operates relays (X) or (Y) depending upon the condition of relay (B). (This can be done with ten springs).

66 4-6

The Design of Switching Circuits Combine and simplify the networks controlling relays (X), (Y), and (Z) as shown below. U possible, the final circuit should be such that relay (C) is equipped with only a single transfer and no other contact. (This can be done with seventeen springs)

C

A

__ITt

C

A

t.____

{1 (Zl~lt~

D

4-7

+..______

-=-

_,J

Design a transfer tree network for four relays, (A), (B), (C), and (D), to indicate each of the following ten combinations by llghtlng a corresponding lamp. For each comblnation glven, the unlisted relays are released, and no lamp must llght for any combination not shown. (A) (A) and (B) (A) and (C)

(A) and (D) (A),(B) and (C) (A) ,(B) and (D) (A),(B),(C) and (D)

(C) (B) and (C) (B)

This design should be executed to distribute the contact spring load as evenly as possible, consistent with the lowest possible number or contact springs. (This can be done with a distribution 11-11-8-3 springs, not necessarily in that order). 4-8

Design a contact network to indicate by lighting a lamp when three and only three out of six relays are operated, with the following two exceptions: relays designated (A), (C), and (E) are never operated simultaneously, and relays (B), (D), and (F) are never operated simultaneously, under normal conditions. The lamp should not llght tr either of these two combinations of operated- relays does occur inadvertently. Draw the required contact network. (This can be done with thirty

4-9

springs).

A circuit includes a group of six relays numbered in order from 1 through 6. Design a contact network which is closed when all relays above a certain one are operated and all others are released. The network should be closed when all relays are operated, but open when all are released. (This can be done with rtrteen springs).

4-10 A circuit includes a group of eight relays numbered from 1 through 8. Design a network which is closed when any two adjacent relays are operated and all others are released. (This can be done with forty springs).

Relay Contact Network Configurations 4 -11 A circuit

67

includes a large number of relays which operate and release In a random fashion, changing condition, however, only one relay at a time. It ls desired to Indicate by means of a ring of lamps whether the number of operated relays Is Increasing or decreasing. Only one lamp should be lighted at a time, and some lamp should be lighted under all conditions of the relays. An increase in operated relays is Indicated by clockwise rotation of lighted lamps; a decrease In operated relays by counterclockwise rotation. Determine the minimum number of lamps necessary to achieve this eHect and design a contact network to control the lamps. (A suitable network for ten relays can be constructed with eighty-one springs).

Chapter 5 SWITCHING ALGEBRA AND MANIPULATION OF RELAY CONTACT NETWORKS

In the design of circuits, rearrangements of contact networks ar~ often necessary either to reduce to a minimum the number of contacts, or to meet requirements as to the number of springs permissible on a certain relay. These network manipulations and simplifications may be performed by methods of inspection similar to those indicated in Chapter 4. However, to carry out the manipulations of complex networks in this manner with any facility, considerable experience is necessary. The designer learns to recognize certain network configurations together with possible rearrangements, and develops his own favorite rules. During the design process the circuit must be redrawn repeatedly to afford a check on the validity of the manipulations employed. The examples quoted in the last chapter indicate that the simplification by inspection may become tedious and time-consuming. In order to expedite relay contact network design and manipulation, an adaptation of the algebra of logic can be used 1 • The algebra of logic, created to introduce mathematical discipline into philosophical logic, is based upon the relationships among two-valued variables. The fundamental similarity between contact networks and prqpositions in logic was mentioned in Chapter 4. This similarity led to the publication by Shannon in 1938 of an algebraic method, developed from the algebra of logic, applicable to relay contact networks2. Switching algebra offers, first, a very convenient notation for expressing contact networks. Utilizing this notation, the algebra then permits manipulation of any two - terminal series - parallel network into a variety of equivalent forms, including the simplest seriesparallel form, with mathematical rigor. Switching algebra is an extremely useful design tool for setting up, simplifying, and combining complex networks. However, in its present form it takes no account of time or sequential relationships, and is of appreciable value only in the later stages of design, after the circuit plan has been clearly formulated. 1• 2

See the references

at the end or this chapter.

- 68 -

Switching Algebra and Manipulation 5.1

of Networks

69

SYMBOLIC NOTATION IN SWITCHING ALGEBRA

In switching algebra as used in this volume, letters, in general, are employed as the symbols to indicate either ·two-terminal contact networks or single contacts of a relay. It is customary to designate a relay with a particular letter and thereafter represent each contact of the relay, and the two-terminal network controlling the relay, by the same letter. A primed letter indicates a contact (or a network) which is the negative of the contact (or network) designated by the same letter unprimed. For convenience, make-contacts are designated by unprimed letters. Finally, a series connection is indicated by a sign of addition, and a parallel connection by a sign of multiplication. As in ordinary algebra, the latter sign is usually omitted. To illustrate: A -t- B' is the notation for a make-contact on relay (A) in series with a break contact on relay (B); AB' indicates a parallel connection of the same contacts.

B A' +B'C

A

C

A

C

B

0

AB' +C'O

At__ B

(A+e>[c'+o (E'+Fl] D C F

E

Fig. 5-1

Simple Networks

and Their Algebraic

Equivalents

70

The Design of Switching

Circuits

The four networks shown in Fig. 5-1 are represented by the expressions A(B + C), A' + B'C, AB' + C'D, and (A+ B) [C' + D(E' + F)j. It can be seen how the use of parentheses and brackets facilitates the expression of the more complicated networks. Any series-parallel twoterminal network may be represented in this fashion. The convenience of such notation is obvious. The algebraic expression representing the conditions for closure of a contact network X is usually spoken of as a function of X, written f(X). Thus, if the conditions for operating relay (R) were relays (A) and (B) operated, the algebraic representation would be: f(R) = A + B. Although a contact on the controlled relay is often included in the network exerting the control, as in the expression for a circuit containing of the expresa lock-up path [f(R) =(A+ B)(R + C)j, the arrangement sion is such that the use of the same letter to represent both a contact and a network should cause no confusion. In numerical algebra the symbols may represent variables, each of which may assume one of a number of values in a particular set of circumstances. This is true also of switching algebra with the limitation that. since contacts and contact networks are two-valued devices, each variable can take on only one of two values, values corresponding either to an open or to a closed path. The quality which this value represents is defined as "hindrance", somewhat analogous to resistance. In the algebrai..J_c_o_N_N_E_cr-,-Ng LEAD

NETWORK xab

A

n

CONNECTING EAD

A1

r'

~ c•l

~1

B1

NETWORK Xm,

G1

':(NECTING LEAD C Fig. 5-25 Graphical Method or Deriving Negative Network (Xab)' = Xmn

Switching Algebra

and Manipulation

~fCx>--AIL -:-

(X)

Fig. 5-26

5.9

of Networks

99

i::r---4 •~

Relationship between Direct and Shunt Relay Control Paths

CONSIDERATION

OF BASIC RELAY OPERATING

PATH

In the design of relay circuits, the designer is not always at liberty to use make and break contacts as he wishes. For example, it may be required to operate a relay (X) when another relay (A) operates, yet only a break contact on (A) is available. This requires the use of a shunt path to operate (X). The basic control paths for a relay were discussed in Chapter 3, and the relationship between direct and shunt control that permits this kind of rearrangement was pointed out. Further discussion of the subject is warranted at this time. A

¾D

t

D

8

~I•~

=

rt

A (X)

8 -:-

8

=

re

(X)

A -:-

=

Fig. 5-27 Converting

a Direct Control Path to Equivalent Forms:

l(X) = A + B

The Design of Switching Circuits

100

8

.!. f'CX)

= A (B+X)= AB+

AX

-:-

Fig. 5-28 Manipilations

_L°e

i__L

of Relay Control Paths:

I(X) = A(B

t 0

-!-°e

A~½

lib_ IX)

+

0

X)

C

J.._

11r:i_

=

-=

-=-

A

I (XI ., A+B IC+Xl

~ -

-

lib_

-=-

e

C

lib_ =

Fig. 5-29

Manlpilations

e

of Relay Control Paths:

C

A

f(X) = A + B(C + X)

-

Switching Algebra

and Manipulation

of Networks

101

If f(X) represents the operating condition for relay (X), the normal control arrangement is to place f(X) in series with the winding of (X). However, the same results, neglecting the time characteristics of the relay, are attained by connecting the negative of f(X), or [ f(X)] ', in shunt with the relay. This is illustrated in Fig. 5-26. If the operating f(X) =A+

path is made up of a sum of factors,

for example:

B

any one or more of the sum factors can be split off from f(X}, and the negative of the split-off factors may be placed in shunt with the relay without affecting the control conditions. The following four arrangements, where subscripts d and s represent direct and shunt paths respectively, are equivalent for control of relay (X): fd(X) =A+

B

fd(X) = A,

f 5 (X) = B'

f 5 (X) =A',

fd(X) = B

f 5 (X) = A'B' The four circuits

are shown on Fig. 5-27.

Since A and B may represent networks instead of individual contacts, the relationship is perfectly general and permits any degree of manipulation, provided only that it is a sum factor that is split off for a shunt path. If f(X) is composed only of product factors that cannot be converted to a sum of products, the rearrangement is not possible unless the status of the relay is reversed, that is, the operated condition is considered normal. In the latter case, the negative of f(X), which is a set of series terms, becomes the direct operating condition. Two additional examples Figs. 5-28 and 5-29.

5.10

of this type of manipulation

RECAPITULATION OF THE POSTULATES OF SWITCHING ALGEBRA

are shown on

AND THEOREMS

The principles and relations used in switching algebra, as discussed in this chapter, are scattered over some thirty-odd pages of text. For the convenience of the reader in review and reference, the postulates and theorems have been tabulated in a concise form on the following page.

The Design of Switching Circuits

102 DEFINITIONS

POSTULATES

Addition

( 1)

X = 0 or X = 1, where X Is a contact or a network

(2a)

O·O • 0

(2b)

1+1= 1

(3a)

1· 1 = 1

(+) • AND = Serles

Multiplication ( •) • OR = Parallel

(3b) 0+0=0

Circuit States

(4a)

0 • Closed Circuit 1 • Open Circuit

1·0 = 0·1 = 0

(4b) 0 + 1.= 1 + 0 = 1

THEOREMS Addition and Multiplication

Negatives,

and O and 1

(la)

X+Y=Y+X

(6a)

(X)' = X'

(lb)

XY = YX

(6b)

(X')' = X

(2a)

(X + Y) + Z = X + (Y + Z)

(7a)

(X + Y + Z ...

)' = X' • Y' • Z' •••

(2b)

(XY) Z = X(YZ)

(7b)

(X. y • Z ...

)' = X' + Y' + Z' ..•

(3a) XY + XZ = X(Y + Z)

(Ba) X' + X = 1

(3b)

(8b)

X'X = 0

(4a) X+X=X

(9a)

0 +X = X

(4b)

XX =X

(9b)

1 •X =X

(5a)

X+XY=X

(X + Y) (X + Z) = X + YZ

(Sb) X(X + Y) = X

Supplementary

(lOa)

1 + X"

(lOb)

0 •X =0

(lla)

(X + Y') Y = XY

(llb)

XY' + Y = X + Y

1

Theorems (12a)

(X + Y) ( X' + Z) (Y + Z) = (X + Y) (X' + Z)

(12b)

XZ + X'Y + YZ = XZ + X'Y

( 13 )

(X + Y) (X' + Z) = XZ + X'Y

(14a)

f(X) = A• f(X) A • , , A' • 0 + A' • f(X) A • o, A' • 1

(14b)

f(X) = [ A + f(X)A • o, A' • 1] [ A' + f(Xl A • l, A' • o]

Switching Algebra

and Manipulation REFERENCES

1.

G. Boole, "The Mathematical tion o( the Laws of Thought" L. Couterat,

"The Algebra

Lewis and Langford, 2.

of Networks

103

FOR CHAPTER 5

Analysis of Logic" (Cambridge, 1847). "An Investiga(London, 1854; reprinted Chicago, 1916). of Logic"

"Symbolic

(Open Court, 1914).

Logic"

C. E. Shannon, "A Symbolic Analysis tions of A.I.E.E., Vol. 57, 1938).

(The Century Co., 1932). of Relay and Switching

Circuits"

(Transac-

G. A. Montgomerie, "Sketch for an Algebra of Relay and Contactor Oournal of I. of E.E., Vol. 95, part m, No. 36, July 1948). J. Riordan Networks"

and C. E. Shannon, "The Number of Two-Terminal Oournal of Math. and Physics, Vol. 21, No. 2, 1942).

C. E. Shannon, "The Synthesis Technical Journal, Vol. XXVIII,

of Two-Terminal Switching No. 1, Jan. 1949).

G. H. Buffery, "A Contribution to the Algebra (Proceedings or I. or E.E., Vol. 97, No. 1950).

Circuits"

Serles-Parallel

Circuits"

(Bell System

or Relay and Switch

Contacts"

PROBLEMSFORCHAPTER5* 5-1

(a) Sketch the contact expressions: 1. A+ B

networks

represented

by the following

switching

algebra

CD

+

2. AB (C • D) 3. (A + B)(C + D) E + F (GH + K)

4. A (BC + D) [ (E + F) G + HK) (b) Write the switching the top or page 104. 5-2

(a) Express

algebra expressions

the following

which represent

the networks

as a sum of products:

A (B + C) [ D + E (F + G) I + HK (b) Express

the above as a product of sums.

(c) Simplify

the following

expressions:

1. (A + B + D)(E + C)(B + C + D)(A + E)

(5 terms)

2. (B + D + E)(A + B + C + D)(A + D + E)

(7 terms)

3. BDE + FA (C + F + GE) + BE

(4 terms)

4. AB + BDE + CA + DCE

(5 terms)

5. ABE+

(5 terms)

6. AB+

BCFG + B (D + G) + BCDF AC+

• In doing the problems

BC for this chapter,

(5 terms) show all steps In arriving

at a solution.

shown at

104

The Design of Switching Circuits

I. o------

...[,

__ A_C

Ja.---0---------0

0

3• ~A-D;J-EG-K•~

BHE C

4

•

H--L J--M

F

1:J-----r

--o

F-G

L~.J •---------___. Networks for put

5-3

(b) of Problem

(a) Sketch the circuits represented by the following spring combinations where possible. 1. A'B + A(B'

~

C)

5-1

expressions,

showing transfer

(8 springs)

2. (A + B + C')(A' + D + E)(F + G + C)

(16 springs)

3. (A'B + B'CD + C'E)(A + E')

(14 springs)

4. F(A(B + C) + D'E] (A'+ D + E')

(15 springs)

(b) Write the negative of each of the above algebraic

expressions.

(c) Sketch the circuits represented by the negative expressions which you obtained In part (b), showing transfer spring combinations where possible. 5-4

Simplify

the following

(a) (A+B+C)(A' (b) ABC

contact networks by algebra.

+B' +C')(A+B'

+C')(A' +B+C)(A+B+C')(A'

1 +A'BC+A'BC'+ABC+A'B'C

(c) (A + B'C)(A'

+ AB'C

+ B' + C)

+B+ C')

(2 terms) (2 terms) (5 terms)

(d) AB (C'D + A' + B) + B' (C + D')(A + B)

(4 terms)

(e) (A+ D')(B'E)

(8 terms)

+ A'C'DF

+ B'C'EF

Switching Algebra

and Manipulation

of Networks

105

5-5

Rearrange the following network to use break-make transfers Instead of the makebreak transfers (continuities) shown, without introducing the hazard of relay (X) releasing during the acting time of either transfer combination.

5-6

Five relays (A), (B), (C), (D), and (E) may be operated in any possible combination. Design a circuit with a minimum number of contact springs which lights a lamp (L) for all combinations except the following: (This can be done with 26 springs.) (A + B + C' + D' + E) (A' + B + C' + D + E) (A + B' + C' + D + E) (A' + B + C + D + E')

(A + B + C + D' + E') (A' + B' + C + D + E) (A + B' + C + D' + E) (A + B + C' + D + E') (A' + B + C + D' + E) (A+ B' + C + D + E') (A' + B + C + D + E') 5-7

Simplify

and draw the circuit

for each of the following

expressions:

(a) ( B' + (D' + F)(E' +A+ G)]( B + F(D + A +E')](A+G+C(D (b) (A'+D'+B)(B+AC)(D+C+A') (c)

( B'CE'(A'

5-8

(6terms,

+ D) + BD +AD' Fl( F(CE' +A'B'

(d) (W + XY')(X'

+ E + F')]( BD'F+DF'] (5 terms, 10 springs)

+ B'D) + B'(C' + E)(A' +D)] (6 terms,

lOsprings) 12 springs)

+ Y') + (Z' + XY')(X + Y) + (Y + w·z)j(w· + Z')+ (X' + W'Z)(W + Z) (8 terms, 12 springs)

Combine each of the following four sets of networks works for operating the specified relays:

Into single multi-terminal

(a) f(X) f(Y) f(Z)

=A+B+C =A+B+D =B+D+E

(12 springs)

(b) f(X) f(Z)

=A'+ B = (A + B + C)(C' + D)(C + E)

(12 springs)

(c) f(W) f(X) f(Y) f(Z)

= = = =

(13 springs)

A+ B' ABD AB+ C (A + B')(A'

+ B)

net-

106

The Design of Switching Circuits (d) ((V) f(W) ((X) f(Y) ((Z)

5-9

5-10

= A' + B' + C' + D' = A' + B + C' + D' = A' + C' + D = A+ C' =C

(12 springs)

Design and combine contact networks to light two lamps under the following tions. U possible, use no more than one transfer per relay. Green Lamp

Red Lamp

A+ B' + C' A'+ B + C' A+B+C A+C B+C

A+ B' + C' A'+ B + C'

Simplify and draw the circuits corresponding to each of the following tables of combi nations. Only the combinations shown in each individual table can be used in the simplification of the corresponding circuit.

A B C D (a)

0 1 1

0 0 0

0 0 1

0 0 0

(c) (8 springs)

~

~

£ Q

1 1 0 0 0

1 0 0 0 1 1

0 0 0 0 0 0

0 0 0 1 1 1

(11 springs)

1 1 1 1 0 0

1 0 0 0 0 1

0 1 0 0 0 0

(9 springs)

1 (b)

0 1 0 0

1 0 0 0

0 0

1 0

0 0 0 1

(d) (16 springs)

1 0 1 0

1 1 5-11

condi-

A circuit designer gathers from several sources the condltlons under which red and green indicating lamps are to be lighted. The conditions are dependent upon the state of three leads, each o( which can be connec.ted to the winding of a relay• Design the simplest possible network on the three relays to control the two lamps. (This can be done with 8 springs.) The red lamp lights when: Input lead b is grounded, lead c not grounded Input lead a is grounded, lead b is grounded Input lead a not grounded, lead c is grounded Input lead c is grounded, lead b not grounded Input lead b Is grounded, lead a not grounded No leads grounded All leads grounded The green lamp lights when: Input lead c Is grounded, lead b not grounded Input lead b not grounded, lead a Is grounded Only lead c is grounded

Switching Algebra

and Manipulation

107

of Networks

5-12 Develop a formal proof, based on the postulates and on theorems (1) through (11), of theorem (12a) and theorem (12b). Work Crom the left side or the theorem to the right, and do not use the method o! perfect induction. 5-13 Design the most economical contact network !or !our relays to be closed when one, three, or !our o! the relays are operated. (This can be done wrth 22 springs.) 5-14 Design a contact network to be applied to a large number o! relays (nine or more) which will be closed when the number or operated relays is not divisible by three. Consider that zero is not divisible by three. (This can be done with 26 transfers on 9 relays.) 5-15 Design a contact network to be applied to a large number of relays (nine or more) which will be closed when the number of operated relays ls not divisible by three. Consider that zero is divisible by three. (This can be done with 22 transfers on 9 relays.) 5-16 A bridge network may be resolved into its equivalent series-parallel form by any of three methods: (1), tracing all possible paths from input to output terminal; (2), determining all means by which the network may be opened; or (3), employing starmesh or Y-Delta transformations. Using one or these methods, resolve the bridge network shown below. Check the solution obtained by carrying out resolution by one or the remaining methods. 0

~

o

A

a

C

B

A

~ 5-17

Prove the truth or the following (A + B')(A'

5-18

Prove the truth

oF~ E

D o

equality

+ B)(A + C) =(A+ of the following

y

J

G

o H o

by the use or theorem

~ (12a):

B')(A' + B)(B + C)

equality

by the use or theorem

(A + B' )(B + C')(C + A') = (A' + B)( B' + C)(C' + A)

(12a):

0

Chapter 6

DESIGN OF COMBINATIONAL RELAY CIRCUITS

6.1

GENERAL DESIGN PROCEDURE

The design procedure for a relay circuit into three parts:

may be divided

roughly

(1) A determination of the requirements of the circuit in terms of what it must accomplish and what is available for the control of the circuit. This will include a determination of the number of input and output leads, the nature of signals or conditions established on these leads, and the time sequence of the various input and output signals. (2) An analysis of the requirements, and the development of a scheme or plan of operation. This will include a determination of the number of relays and a plan of time sequence for the action and interaction of these relays. (3) The detailed design of contact networks and circuit arrangements for receiving the input signals, controlling the relays, and establishing the desired output conditions in accordance with the plan. During the design of a practical circuit there is often considerable overlapping of the work classified under these three headings. The work under (2) and (3) may indicate that different input conditions are necessary or desirable, while the detailed work under (3) often indicates modifications of the plan developed under (2) which will produce a more satisfactory circuit. Because of the fact that in the design of switching or control circuits there are always a number of alternative ways in which a particular result may be accomplished, experience, good judgment, and ingenuity are necessary factors in the design of such circuits. In this and the following chapter, design procedures for the basic classes of relay circuits will be outlined in some detail. Emphasis will be placed on the work indicated in (2) and (3) above; the work of part (1) will be assumed completed, since it is often determined by factors over which the designer has little control, or by engineering judgments

- 108 -

Combinational

Relay Circuits

109

based on knowledge of the general requirements and capabilities of particular apparatus.

6.2

INPUT AND OUTPUT

of a particular

system

CONDITIONS

As defined in Chapter 3, the term "relay circuit" is an elastic term used to group together those relays that may be considered as a unit. By its nature, a relay circuit must be controlled and must exercise control over leads connected to other circuits and apparatus. Usually, but not always, the incoming control conditions are carried by leads separate from those carrying the output conditions. The nature and form of input control conditions which transmit information to a circuit composed of relays or other two-valued devices are necessarily aff ecfed by the characteristics of the elemental apparatus. In general, the basic input (and output) signals are two-valued also. Depending upon the type of receptor apparatus, the signals may be presence or absence of ground on a lead, pulse or no pulse, positive or negative voltage or current, etc. Under certain circumstances, the number of signal values on a single input path may be made greater than two. For example, the marginal and sensitive relay arrangement shown in Chapter 3 on Fig. 3-12 can respond to three conditions over a single path. However, this can be considered a special case. Although these narrow limits to the type of usable input signals may seem unduly restrictive, there are two mitigating aspects to the situation. The first is that a limited number of two-valued signals, taken in combinations or in sequence, can provide a much higher number of distinctive input conditions. The effect of sequence is to increase the number of overall input conditions far beyond the 2n derivable from n individual conditions taken in combination. This aspect permits imposing upon a circuit an almost unlimited quantity of input information via a finite number of input paths. The second important aspect of input control is that if the original information is not in two-valued form, conversion apparatus can almost always be interposed between the sources of raw information and the control or switching circuits in order to change the information into an appropriate form. Many well-known examples of this type of conversion can be quoted. A thermostat converts gradual temperature changes into circuit opens or closures; float and lever systems close or open circuit paths With a change in liquid level; and so on. With this understood, it will be assumed in this text that control or input signals are always available in proper form to activate the circuits to be designed; that is, the signals will be two-valued. The nature and design of any conversion apparatus that may be required in

The Design of Switching

110 actual practice will scope of this text.

be considered

a separate

subject

Circuits

not within

the

Output co~ditions are of the same nature as input conditions with the distinction that the input conditions initiate and control action in the circuit under consideration, while the output conditions either control other circuits or deliver information.

6.3

GENERAL TYPES OF RELAY CIRCUITS

Relay circuits can be divided into two general types, designated "combinational" and "sequential" circuits. In the combinational circuit, a combination of input conditions establishes a definite combination of output conditions independent of the order in which the input conditions are applied. In the sequential circuit, the time sequence in which input conditions are applied affects the output conditions established. A factor of memory is always contained in a sequential circuit. first

6.4

This chapter will be devoted to the design fundamentals or combinational type of relay circuit.

CAPABILITIES

OF COMBINATIONAL

of the

CIRCUITS

In a typical combinational circuit, each of n input leads connects to the winding of a relay of the circuit. The relays are controlled by signals on the leads. Networks of contacts on the relays control the signals (ground or open) on the output leads. Since the input conditions are two-valued, there are 2 n possible combinations of input conditions (input combinations), and then relays can operate in 2" corresponding combinations. The contact networks can establish a combination of conditions on the output leads (output combinations) for each of the 2 n input combinations so that the maximum possible number of unique output circuit, a particular input combinations is also 2 "· In a combinational combination will always produce the same output combination regardless of the sequence in which events occur on the input leads. However, the circuit may be arranged so that several different input combinations produce the same output combination. In many circuits, all the possible input combinations do not occur in normal operation. In any case the number of different output combinations is never greater than the number of input combinations. The preceding discussion assumed that each input lead connects to a relay which serves to detect signals on this lead. In particular cases, it is frequently possible to eliminate this relay and let the lead connect directly into the contact network. The purpose of the relay is to provide contacts which, in effect, repeat the input signal (and its

Combinational

Relay Circuits

111

negative) in as many different circuit paths as are necessary to achieve the required results. When an input signal condition (e.g., ground) is the same as one or more output conditions, it may be possible to arrange the contact network so that the input signal-detecting relay requires only a single make-contact closing a path to ground. Hence the relay may be omitted, and the signal on the input lead substituted for that produced by the relay contact. If the circuit is considered from the output end, the number of required output combinations is often a determining factor in the number of relays in the circuit. For example, if twenty output combinations are reQuired, at least five relays, and hence five inputs, will be needed. In another case, if a circuit is required to start from an inactive condition and provide eight active output combinations, three relays with their maximum of eight combinations would not suffice.

Consider now the individualoutputleads of acombinationalcircuit. If each lead is grounded for only one input combination, there can be a maximum of 2 n different output leads, each lead corresponding to a particular input combination. However, an output lead may be grounded for any number of the 2n input combinations. For example, a circuit with two input leads may have 22 = 4 input combinations. An output lead may be grounded for any combination of these four input combinations. This gives a total of 2 4 = 16 networks to which output leads may be connected. Two of these 16 ways - the one which produces a grounded output for all input combinations (a permanently grounded output lead), and the one which produces a grounded output for none of the input combinations (an open output lead) - are trivial and can be omitted from consideration. Thus, in a circuit with n input signals, there are (2 2 n -2). useful ways in which output leads may be connected through contacts of the input relays.

6.5

DESIGN PROCEDURE

FOR COMBINATIONAL

CIRCUITS

The technique of combinational circuit design entails consideration of each output lead separately, and determination of the input combination which will establish the desired output condition on this lead; that is, the circuit is synthesized from the output end. Usually it is convenient first to develop a separate contact network for each output lead and later to combine the networks into a single multi-terminal network. The circuit for each individual output lead may be designed by inspection or by formal methods. To develop a circuit by inspection of requirements, the entire group of input and output combinations is examined for coincidences of input and output signals, symmetry of requirements, or other features

The Design of Switching

112

Circuits

of the requirements as a whole which will indicate significant relations between inputs and outputs. The object of this inspection is to obtain directly a statement describing the relations among input signals which must cause a given output signal. These statements are determined for each output and the corresponding circuit paths are combined to form the required network.

In the formal method, circuits are obtained by following routine procedures. To obtain a circuit for a given output lead with this method, every possible input combination is examined to determine which combinations correspond to closed circuit conditions of a particular output lead. For each of the combinations, an individual output circuit path, closed only for that combination, is developed. This path will consist of a series circuit containing one contact on every input relay of the circuit, the contact being a make or a break respectively for relays operated or released in the combination. The individual paths are then combined and simplified to give the circuit for one output lead. This is repeated for. each output lead and, finally, the circuits for the several outputs are combined where possible. It is likely that all the possible input combinations do not occur in normal operation, particularly when there are more than four or five individual input signals. In these circuits there is often considerable latitude in manipulating the contact network, since the designer may elect to let those combinations which are never expected to occur correspond to either open or closed conditions of the network being designed. This often permits considerable simplification of the network, as will be illustrated in the next section. The normally unused combinations, however, represent trouble conditions if they occur during the circuit action. Where closure of output leads during such a trouble condition may cause serious external reactions, a strict requirement may be placed on the circuit that ail output leads must be closed only for specifically defined input combinations and must be opened for all others. Another factor that may affect the final form of a combinational circuit is the effect of permitting output leads to be connected together, through contacts of the circuit, during combinations which provide for no output signals on the leads. A circuit which provides complete independence among output leads often requires more contacts than a circuit which does not provide this independence.

6.6

EXAMPLES

OF COMBINATIONAL

CIRCUIT DESIGN

Combinational circuits are designed to activate a particular output lead for a single input combination or a multiplicity of input combinations, or to interconnect two or more output leads for single or multiple input combinations. The basic principles of design will be

Combinational

Relay Circuits

113 A'+ e'+c'+o' I

t

I

0

l

t

:l

A1+ 81+G+0 1

o I

1 L...o A1+ e'+c +o

o

1+0 1 I rA'+8+C I l__,A'+e+c'+o

lA

+8+C+0

1

1

0

L...o A1+ e +c +o

A +e 1+c 1+0 1

-=-

L...oA+e'+c'+o

Fig. 6-1

A Fully-Developed

Four-Relay

Tree

illustrated by a consideration of the simplest case, a circuit where each output corresponds to a single input combination. Several circuits, based on the same primary requirements, will be designed to indicate the effect of modifying fringe requirements. Following this presentation, more elaborate design examples will be discussed. A familiar example of a combinational circuit with a single output per input combination is the "transfer tree" which was discussed in Chapter 4 and is reproduced here as Fig. 6-1. Input signals appear on the operating leads for relays (A), (B), (C), and (D). For four relays there are 24 or 16 possible relay combinations, and the 16 individual output combinations corresponding to these relay combinations appear on 16 output leads. No output corresponds to more than one relay combination, and every output path through the network includes a contact on each relay.

The transfer tree is the optimum circuit form when a separate output is required for each of every possible input combination. When

114

The Design of Switching

Circuits

outputs correspond to single input conditions, but not all input conditions are comprehended in the requirements, the circuit form may or may not be a partial tree, depending upon subsidiary requirements. In the example below, assume that the table of requirements must be followed exactly and that output leads must be independent. The requirements are as follows: Input Leads Grounded

Output Leads Grounded

b

s

C

a.c

u

a,b.c

V

b.c

none

a,b

none

a

none

none

none

To illustrate the formal method of circuit design, the algebraic equation for each of the above output conditions is written from a consideration of the requirements, assuming a relay corresponding to each input lead: f(s)

= A'+ B + C'

f(t)

= A'+ B' + C

f(u) = A+ B' + C f(v)

(6-1)

= A+B+C

These equations•, and their corresponding contact networks, may be combined by inspection to give the circuit of Fig. 6-2. The resultant circuit is a partially developed tree, each output being disjunctive with respect to any other.

If the requirements are such that some input combinations may be assumed never to occur in normal operation (or that the corresponding output conditions are inconsequential if they do occur), a simplification of the basic tree is possible. These combinations, which will be called "invalid" combinations. may be used optionally to eliminate contacts in • Throughout this text, pertinent equations or groups of equations are identified by hyphenated numbers. The first part or the number Indicates the chapter in which the equation Is first given, and the second part shows the number within the chapter. The theorems and postulates which were given in Chapter 5, however, are numbered consecutively without reference to the chapter.

Combinational

Relay Circuits

115 (Bl

Fig. 6-2 A Partially-Developed

Tree

(Equations

6-1)

the circuit. The tabular notation and method of simplification introduced in Chapter 5 offer a direct approach to the simplification of the network of such circuit. For illustration, let the requirements satisfied by the circuit of Fig. 6-2 be modified by stating that the input combinations (b,c), (a,b), and (a) are invalid. (The condition of "no inputs grounded" will occur as the normal condition when the circuit is idle.) However, it will still be required that the outputs be independent, or disjunctive. The table is then drawn up as follows:

A B C

Output Lead Grounded

1.

1

0

1

s

2.

1

1

0

t

3.

0

1

0

u

4.

0

0

0

V

5.

1

0

0

6.

0

0

1

7.

0

1

1

Invalid combinations

Since the invalid combinations represented by lines 5, 6, and 7 of the above table never occur, circuit paths represented by these combinations will never be closed and can be used for contact network simplification in the following manner. Considering the network for output lead s, this network can be closed to ground not only for the active condition represented by line 1, but also may include paths for the invalid conditions of lines 5, 6, and 7. Taking lines 1 and 5 or lines 1 and 6 together, then, it is evident that either C' or A' can be eliminated, with the result that f(s) = A' + B, or B + C'. Considering lines 2 and 5, B can be eliminated from the expression for f(t). Similarly, by combining lines 3 and 7, and lines 4 and 5 or lines 4 and 6, f(u) and f(v) can be simplified, respectively. The algebraic notations for grounding the output leads then become:

116

The Design of Switching f(s)

= A' + B; or B + C'

f(t)

= A' + C

f(u)

= A+ B'

Circuits

(6-2)

f(v) = A + B; or B + C Comparison of equations (6-2) with equations (6-1) indicates of the invalid combinations.

the utility

Combining paths and grouping contacts with their relay windings gives the circuit of Fig. 6-3A. However, in retrospect it would appear that the network for f(t} has been simplified to a point where it cannot be easily combined with the other paths. If f(t) is left unsimplified, f(t) = A' + B' + C; it can be combined conveniently with f(s) = A' + B and the further simplification of Fig. 6-3B results. The point at which advantageous simplification becomes over-simplification depends upon each individual case; only experience can teach the designer to recognize when it has been reached. It should be noted that,in the circuit of Fig.6-3B, the output paths have been kept disjunctive; that is, no two output paths can be connected together.

12 SPRINGS

II SPRINGS

A

B

Flg. 6-3 A Modified Tree

(Equations 6-2)

(8)

------------

s

a------,t~-~---t----.11..1-------------u

(Cl C

-------------------A+-'Flg. 6-4 Non-Disjunctive

L------t

,r-H-~ Circuits

(Equations 6-2)

Combinational

Relay Circuits

117

on the circuit is removed, that is, if If, now, the last restriction the output leads need not be disjunctive, another step of simplification becomes possible. This can be determined by examining equations (6-2) again. It will be noted that the conditions for grounding outputs sand tare

f{s) = A' + B f(t)

= A' + C

which, in this case, tact since B and C sponding circuit is put lead v grounded and t are connected fied requirements.

permits combining are never validly shown on Fig. 6-4. when relays A, B, together. However,

the two A' factors into one conactivated without A. The correWith this circuit, not only is outand C are operated, but paths s this is permitted by the modi-

B

A Fig. 6-5 Circuit

with Relay Eliminated

One more circuit form to meet the basic requirements is possible. It was stated earlier in this chapter that, if output conditions are of the same nature as input conditions and if a circuit relay can be reduced to a single make contact which repeats the corresponding input condition, that circuit relay may be eliminated. By judicious choice of output expressions from equations (6-2) and by a suitable arrangement of contacts, the circuit configuration of Fig. 6- 5A can be attained. This can be changed to the circuit of Fig. 6-5B which employs only two relays. Four circuits have now been designed to meet the same basic set of requirements. It should be clear that other equivalent circuits are possible. Each of the four designed circuits grounds the required output lead for the appropriate set of input conditions, but the circuits differ in other respects. These differences are tabulated at the top of the next page.

The Design of Switching Circuits

118

Clrcult

Fig. 6-2

Circuit

Apparatus 3 relays

15 springs

Characteristics

No output for invalid input combination. Output paths disjunctlve.

Fig. 6-3B

3 relays 11 springs

Outputs grounded for Invalid Input combinations. v). (Example: a,bOutput paths disjunctlve.

Fig. 6-4

3 relays 10 springs

Outputs grounded for Invalid Input combinations. s,t). (Example: b,cOutput paths not disjunctive: (a,b,c grounds v, connects s, t together).

Fig. 6-5B

2 relays 10 springs

Outputs grounded for invalid Input combinations. Output paths not disjunctive. One input condition used as an output condition.

This development of a variety of circuits to perform the same basic requirements illustrates a common principle in circuit design: in general, as requirements become more rigid, the circuit becomes more complex and requires more apparatus. The specific requirements and the general situation surrounding a design problem must always be scrutinized carefully to determine in what manner the circuit should be designed and what short-cuts may be taken. Also, a practical consideration that may affect the degree to which combination and simplification can be carried is the electrical load capabilities of the relay contacts. A frequently used type of combinational circuit is one in which several output leads are grounded in combinations. To obtain a circuit to fit these requirements, a relay is provided, as before, for each input condition. The paths by which each output lead is grounded may be written algebraically, utilizing a contact on every relay. All paths for each output are then placed in parallel and the contact network is combined and simplified. Finally, the composite circuit for all outputs is combined and simplified to whatever extent is possible. For example, the following is required: Input Leads Grounded 1.

a

Output Leads Grounded w,z

2.

b

w.x

3.

C

w,y

4.

a,b

x,z

5.

a.c

y,z

6.

b,c

x,y

7.

none

none

8.

a.b,c

none

Combinational

Relay Circuits

119

It is a requirement that, although lines 7 and 8 in this table of requirements represent combinations which have no corresponding outputs, do occur and hence cannot be used in simplithese jnput combinations fying the networks for the output conditions. Let it also be a stated requirement that no output lead is to be connected to another for any output combination except that for which both are grounded. The contact network paths may then be written as below, although the tabular notation would do just as well. f(w)

= (A+

f(x)

= (A' + B + C') (A+

f(y)

= (A' + B' + C) (A + B' + C) (A' + B + C)

f(z)

= (A+

Simplifying

B' + C') (A' + B + C') (A' + B' + C)

B' + C') (A+

the expression

f(w)

= (A+

f(x)

= B + A'C'

f(y)

= C + A'B'

f(z)

= A+ B'C'

The network in Fig. 6-6.

[A'+

B + C') (A+ B' + C)

(B + C') (B' + C)] (6-4)

paths corresponding

to the equations

(6-4)

are illustrated

of the contact network can be obtained by for f(w) so that it may be combined with the

A

--c-:_J----------z :

8 -{

:J..___-

--c-[::J----Y Fig. 6-6

(6-3)

for each function:

B' + C')

Some simplification rearranging the expression

B + C') (A' + B + C)

Representation

of Equations (6-4)

X

The Design of Switching

120

Circuits

paths for other output leads. For example, f(w) may be manipulated with algebraically to contain elements B'C' and A'C' for combination f(z) and f(x). Carrying out this manipulation: f(w)

= [ (B + C') (B' + C) + A'] = (B'C' +BC+

[ C' + B' + A] (6-5)

A') (A'C' + B' + A)

The resultant network path diagram is shown in Fig. 6-7. Manipulations of this type are recognized only after careful study of all circuit paths, and the combined network must be carefully checked for sneak paths and for disjunctivity when required. Note, in Fig. 6-7, that the presence of A + A' in the path between Z and W, and B' + B in the path between W and X, insures that the outputs are disjunctive. This final circuit is, of course, not the only one which might be derived, since the algebraic equations might be manipulated in some other manner.

e' c'

-=-

I

z

A

:J--· ·--C:',J 8_] C

A'

e'

c'

8

X

y

Fig. 6-7 A Combined and Simplllled

Arrangement

o( Fig. 6-6

Recalling the theorem of switching algebra stating that X + Y = X + X' Y, there is a temptation to obtain an even more simplified network by expanding the equations (6-4) as follows: f(x)

= A'C' + B = A'B'C'

+ B

f(y)

= A'B' + C = A'B'C'

+C

f(z)

= B'C' +A=

+A

f(w)

c:

A'B'C'

(6-6)

(A+ B' + C') (A' + B + C') (A' + B' + C)

Then, since f(x), f(y), and f(z) all contain A'B'C', the three are com})iried and the network paths appear as in Fig. 6-8. Note that the expression for f(w) represents a symmetric one-out-of-three circuit. However, inspection of Fig. 6-8 shows that the output paths are not disjunctive as required, since when relays (A), (B), and (C) are operated, output leads x, y, and z are connected together but not to ground. If the original requirements can be modified to allow this circuit condition to exist, the circuit of Fig. 6-8 represents a saving over

Combinational

Relay Circuits

121

A

A'-{

::=:t=: --------x --------Y

Fig. 6-8 Combining Equations (6-6)

that of Fig. 6-7. Unfortunately, there is no easily applied rule for determining whether or not the network paths are remaining disjunctive • during the simplification of a combinational circuit. The final circuit must be inspected for paths not permitted by the original requirements.

As was the case with the first example taken up in this section, a further simplification will usually result if invalid input combinations never occur or, if they do, any output condition is permissible. If the requirements for this problem are amended to allow the input (a,b,c) to establish any output condition when it occurs, the resultant algebraic expression can be written as follows, with the changes underscored: f(w)

= (A+

f(x)

= (A' + B + C') (A+

= f(y)

B' + C') (A' + B + C') (A' + B' + C) B + C') (A' + B + C) (A+ B + C)

B

= (A' + B' + C) (A' + B + C) (A + B' + C) (A + B + C)

(6-7)

= C f(z)

= (A+

B' + C') (A+ B + C') (A+ B' + C) (A+ B + C)

= A The arrangement of paths is shown in Fig. 6-9. The "no input" condition is reserved for a circuit "normal" condition when outputs must not be grounded. 1 A---8

e=1--c'

A'-[

}-w

e'----c

----A-------------z ----8

X

'----c-------------Y Fig. 5.:9 Modillcatlon

of Fig. 6-8 (Equations

(6-7)

The Design of Switching

122

Circuits

The design of circuits by inspectionof requirements can be illustrated by this problem. The method, as previously stated, is to examine the requirements and determine directly the relations between inputs which must cause each output to be closed. Referring to the table of requirements for the present problem, and assuming a relay operating from each input, it is observed that output lead w must be grounded when exactly one of the relays (A), (B), and (C) is operated. An obvious circuit for f(w), therefore, is a one-out-of-three symmetric circuit. The statement for grounding output lead z is, "(A) and, (B) or (C) is released", which may be written:

is operated

f(z) = A + B'C' The part of the statement: "(B) or (C) is released", represented by B'C', is necessary to insure that z is not grounded when all relays are operated. Similar statements can be made for f(x) and f(y), and the resulting circuit is as shown in Fig. 6-6, with the circuit for f(w) changed to the symmetric form. • A simplified form of the network may be obtained by observing that, although output z is usually grounded when (A) operates, x when (B) operates, and y when (C) operates, none of the three outputs is grounded when all three relays are operated. The network A'B'C', which is open only when all three of the input relays are operated, can then be made common to the paths for the x, y, and z outputs to obtain the circuit of Fig. 6-8, where f(w) is in the symmetric form previously obtained. Inspection of this circuit shows that the outputs are not disjunctive, which may or may not be satisfactory.

If it is assumed that the condition of all inputs grounded simultaneously never occurs (that is, input combination a, b, c is invalid) then the statement, "z is grounded when (A) operates" is sufficient for f(z), and a single contact on (A) will suffice. Similar statements for f(w) and f(x) result in the circuit shown in Fig. 6-9. The choice between the inspection method and the formal method in solving a problem can often be determined by a preliminary analysis of requirements. Inspection often gives a simplified solution directly with less circuit manipulation, but final circuits must be carefully analyzed to see that they conform to requirements, maintain the necessary disjunctivity, and do not contain sneak paths. The type of combinational circuit requiring the most contacts is one whose output consists of connecting separate leads together, rather than of grounding these leads. Simplification is sometimes possible, but the necessity of keeping paths separate requires many contacts. As an example, take the following requirements:

Combinational

123

Relay Circuits

f(s-t)

= A' + B'

f(u-v)

= A + B'

f(w-x)

= A' + B

f(y-z)

=A + B

Input Leads Grounded

Output Conditions

none

connect lead s to t

a

connect lead u to v

b

connect lead w to x

a,b

connect lead y to z

(6-8)

No paths can be combined,

and the circuit

s------r:>-n

of Fig. 6-10 is obtained.

....

Y

-"'--------z

•

•

a~-l·ri-rtl•CJ, b-------------'·

Fig. 6-10 Connections as Output Conditions

Another

(Equations 6-8)

example is the following: Input Leads Grounded a

Connect lead s to t

b

Connect lead t to u

a,b

Connect lead s to u

none

The equations for this circuit f(s-t)

= A + B'

f(t-u)

= A' + B

f(s-u)

=A+

B

Output Conditions

no connections

are:

(6-9)

The Design of Switching Circuits

124

Here certain inter-relationships among the inputs and outputs indicate that simplification may be possible. For instance, the lead s is associated with A, and the lead u with B. In the noncombined form, the circuit is as shown in Fig. 6-llA; by inspection it can be combined into careful the form of Fig. 6-llB. Every circuit of this type requires analysis to determine what simplifications, if any, may be permitted by the requirements and the configuration of the individual networks. s

--ci.-,.,.......,,.

,---t

t --cCH++-__.t

"---u

A

B

Fig. 6-11 Combined Circuit

for Connecting Output Paths

(Equations

6-9)

When designing circuits using more than four relays, the formal method applied to the problem as a whole often becomes cumbersome and the method of inspection is more satisfactory. For example, take the requirements in the table below: Input Leads Grounded

An assumption

Output Leads Grounded

v,w

a

v,x

b

w,x

a,b

v,y

C

w,y

a,c

x,y

b,c

v,z

a,b,c

w,z

d

x,z

a,d

y,z

b,d

none

none

is that no other input conditions will ever occur.

The correspondence between inputs and outputs can be used to facilitate the design of the contact network as follows: (1) Output lead a is grounded whenever input lead w or z is grounded, except for the input combinations (w,z) and (y ,z). This can be stated in terms of relays controlled directly by the input leads

Combinational

125

Relay Circuits

as: output lead a is grounded when relays (W) or (Z) are operaated and not when (W) and (Z) are operated and not when (Y) and (Z) are operated. Or algebraically: f(a) = WZ + (W + Z)' + (Z + Y)' = WZ + W'Z' + Z'Y'

(6-10)

(2) Lead b is grounded when (X) or (Z) is operated and not when (X) and (Z) are operated and not when (Z) and (W) are operated; that is: f(b) = XZ + (X + Z)' + (Z + W)' = XZ + X'Z'

(6-11)

+ Z'W'

(3) Lead c is grounded when (Y) is operated and not when (Y) and (Z) are operated; or lead c is grounded when (V) and (Z) are operated; that is: f(c) = [ Y + (Y + Z)'

J (V

+ Z) (6-12)

= (V + Z) (Y + Z')

(4) Lead d is grounded when (Z) is operated and not when (V) and (Z) are operated; that is: f(d) = Z + (V + Z)' = Z + V' The combined

circuit

(6-13)

is shown in Fig. 6-12.

--b H-i.

l+i (WI ..,.

(Z)

..,.

v--+----t----' w )(------------+-------ti--'

y-----+--------+---' z

._ _______

Fig. 6-12 Five-Relay (Equations

Combinational Circuit 6-10, 6-11, 6-12, and 6-13)

C

d

The Design of Switching

126

Circuits

This network could have been obtained by making a table of the thirty-two possible input combinations including the twenty-one invalid combinations and the eleven valid combinations giving particular output conditions, and combining to eliminate contacts. The work involved by this tabular method would be much longer. Circuit design by inspection of requirements is usually the better method when less than half of the possible input combinations occur in normal operation and the circuit action for invalid combinations is not specified.

1f the requirements of the above example are made more rigid and state that no output leads are to be grounded should the invalid input combinations occur in this particular case, advantage can be taken of the symmetry of the working input conditions. The· inputs are in pairs, and all possible pairs are used. By supplying ground to the contact network developed above through a symmetric circuit closed when two and only two of the five relays are operated, the more rigid requirements are met. This is shown in block form in Fig. 6-13. I I

-1-

2 OUT Of" ~ SYMMETRIC CIRCUIT

I I I

I I I

t (al

I I

0

f (b)

I I

b

f ( c)

f (d)

I

I I I

C

d

Fig. 6-13 Block Diagram of Circuit

(or Removing Ground from Unused Output Combinations or Fig. 6-12

A specituized type of combinational circuit is that in which the requirements are stated in terms of the relative position or serial number of the input leads, rather than in terms of correspondence between specific input conditions and outputs. It can easily be seen that the attempt to cope with problems of this nature by the formal method discussed in this chapter will meet with considerable difficulty. However, there are two profitable methods of attack. Since this type of circuit, in general, utilizes a reiterative network, the technique for designing such circuits described in Chapter 4 is applicable. Another attack, in many cases more direct, is essentially the method of inspection as described in this chapter. The requirements are analyzed carefully to discern some one of a variety of equivalent statements of conditions which is most susceptible to expression in circuit form. 1f such a statement can be found, the circuit is written in algebraic or contact form and manipulated to give the simplest network. Consider,

Combinational Relay Circuits

127

for example, the following positional circuit requirements: Six control leads, La, Lb,Lc, Ld, Le, and Lr are grounded at random, singly and in combination. A single output lead, t, is to be grounded only when La or a number of adjacent control leads starting with lead La are grounded. In the design of a suitable circuit, a relay per control lead is first assumed and the relays are designated (A), (B), (C), (D), (E), and (F) to correspond to the control leads. Given these relays, it should be possible to write, in terms of the relays, the algebraic expression for grounding output t. However, the statement "lead t is grounded only when one or a number of adjacent relays are operated starting with relay (A)" does not appear to lend itself directly to translation into algebraic form. An attempt, then, is made to restate the conditions for network closure in some more suitable form. One such restatement is as follows: "the network should be open when all relays are released or when any relay is released and the succeeding relay is operated; under all other conditions the network should be closed." The algebraic expression corresponding to these requirements for opening a circuit is the negative of that corresponding to requirements for network closure. Writing the negative expression, then: [f(t)]'

=

(A' + B' + C' + D' + F')(A' + B)(B' + C) • (C' + D)(D' + E)(E' + F).

Taking the negative of both sides of this expression, for grounding lead l is obtained.

(6-14) the equation

f(t) = [(A' + ~• + C' + D' + E' + F')(A' + B) • (B' + C)(C' + D)(D' + E)(E' + F)]'

=ABCDEF

+ AB' + BC' + CD' + DE' + EF'

(6-15)

This equation may be reduced to f(t)

=A + AB'

+ BC' + CD' + DE' + EF'

= A + BC' + CD' + DE' + EF'

(6-16)

The network represented by this last expression may be drawn by inspection as shown in symbolic form in Fig. 6-14A, on the next page .. Placing this network on the relays controlled by the leads "L ", a circuit is obtained which fulfills the original requirements. This final circuit is shown in Fig. 6-14B. 6.7

TEMPORARY FALSE OUTPUT CONDITIONS

In obtaining solutions to circuit problems so far, it has been as sumed that a relay closed instantly when energized. Actually, of course,

The Design of Switching Circuits

128

A

B

Fig. 6-14 A Combinational Circuit

Employing a Positional

Network

(Equation

6-16)

there is always some time delay between the instant of closing the relay's operating path and the instant of the relay closing its make contacts or opening its break contacts. Furthermore, there may be an appreciable stagger time among the operation of several contacts on the same relay. In addition, one relay, due to manufacturing tolerances and variations, may have appreciably different operate and release times from another similar relay. It is probable, then, that in a combinational relay circuit the wrong combination of relays will be operated for a brief time after the proper combination has been energized. Usually this time is so brief that it can be tolerated, but the designer should always be aware that it may exist and make sure that the effects of the "race condition", as this is called, are properly provided for. As an illustration of how a wrong output combination occurs momentarily in a race condition, take the input combination where output lead v is to be grounded in Fig. 6-4. Suppose that output lead u is the control lead for a fast-acting relay in another circuit. Input leads a and b are grounded at the same time; relay (A), being lightly loaded, operates in a shorter time than relay (B) and grounds lead u. A short time later relay (B) operates, opening u and grounding lead v; but since u controls a fast-acting relay, there is a distinct possibility that this relay has already operated falsely. Methods of guarding against such a race condition when the output controls a fast-acting relay, are beyond the scope of this chapter, but will be discussed more fully in Chapter 8.

PROBLEMS FOR CHAPTER 6 6-1

Draw the fourteen di1ferent networks using contacts on one or two relays, as discussed in Section 6.4. Not Included are the always open and always closed paths.

Combinational

129

Relay Circuits

6-2

Eight relays are divided Into two equal groups, A and B. or four relays each. The relays or group A are designated (Al), (A2). (A3). (A4), while those or group B are designated (Bl), (B2), (83), (B4). Design a contact network to light a lamp when the designation number of an operated relay In group A Is greater than the designation number of an operated relay in group B. The lamp should light only when one and only one relay in each group is operated. (This network can be designed with less than 35 contact springs.

6-3

Eight relays are divided into two equal groups. A and B, or four relays each. Design a contact network to light one or three lamps under each or the following three conditions: (This can be done with 24 transfers).

6-4

Lamp (Ll):

when more A relays are operated than B relays.

Lamp (L2):

when an equal number of A and B relays are operated.

Lamp (L3):

when more B relays are operated than A relays.

A circuit following

for four relays (A), (B), (C), and (D) operates chart is satisfied:

(a) Design occur.

released

X

(A),(C); all others

released

z

(A) ,(D); all others

released

y,z

(B) .(C); all others released

x,z

(B).(D); all others released

y

(C) ,(D); all others released

x,y

None

None

the required

circuit

that the

Output Leads Grounded

Relays Operated (A) .(B); all others

In such a manner

assuming

that only the combinations

(b) Design a second circuit assuming that all possible combinations outputs should appear only on the above combinations.

above can (13 springs)

can occur, but (31 springs)

Leads can be connected together while ungrounded. 6-5

A circuit following

for four relays (A). (B). (C). and (D) operates chart is satisfied:

(A).(B); all others released

x.y

(A) .(C); all others released

X,Z

(A),(D); all others released

y

(B) ,(C); all others released

z

(B) ,(D); all others

X

released

(B) ,(C) ,(D); all others the required

circuit

that the

Output Leads Grounded

Relays Operated

(a) Design occur.

In such a manner

released

assuming

that

y .z only the combinations

above can (13 springs)

130

The Design of Switching (b) Design a second circuit assuming that all possible combinations out~ts should appear only on the above combinations.

Circuits

can occur. but (29 springs)

Leads can be connected together while ungrounded. 6-6

Three relays (A). (B). and (C). which may operate in any combination. control a red lamp, R. and a green lamp, G. as follows: when all three relays are operated, the red lamp only Is lighted; and when relays (A) and (B) are released and (C) is operated. only the green lamp is lighted. For all other combinations or the relays both lamps are lighted except that no lights are to be operated when all relays are (9 springs) released. Derive the circuit using as rew contacts as possible.

6-7

The table below indicates the requirements (or a circuit which connects together certain leads depending upon the operated combinations or relays (A). (B). (C), and (D). Design the circuit so that when any relay combination not included in the table (15 springs) occurs, no leads are connected together. Relays Operated

Leads Connected Together

(B)

- (s to l). (v to x)

(C)

- (w to x)

(D)

- (u to t)

(A) ,(B)

- (s to u)

(A) ,(D)

- (u to l)

(B) ,(C)

- (v to w), (x to v), (x to w)

(B) .(D)

- (s to l), (u to s). (t to u), (v to x)

(C) ,(D)

- (w to x)

(A) .(B) ,(C)

- (s to u), (v to w)

(A) .(B) ,(D)

- (s to u), (t to u), (s to t)

(B) ,(C) ,(D)

- (s to u), (v to x), (v to w), (w to x)

(A) ,(B) ,{C) ,(D) 6-8

- (s to u), (v to w)

In a typical relay storage unit for binary numbers, the binary digit "'l" may be indicated by an operated relay and the binary digit ·•o·•,by the relay unoperated. Relays (A), (B), (C), and (D) are elements in such a storage unit, where (D) represents 2°, (C) represents 21, etc. Since there are four relays, the unit may be used to hold the binary equivalent to any decimal number from 0 to 15. For example. the decimal number 13 would be stored by operating relays (A), (B), and (D), since the binary equivalent Is 1101. Design contact networks controlled by relays {A), (B), (C), and (D) to be closed each of the following conditions: (a) The stored binary number Is equivalent to an odd decimal

for

number.

{b) The stored binary number ls equivalent to a decimal Including zero.

number divisible

by 4, not

(c) The stored binary number Is equivalent to a decimal Including zero.

number divisible

by 3. not

{d) The stored above 10.

binary

number

Is equivalent

to a decimal

number

below

5 or

Combine and simplify each contact network as far as possible. (The four networks can be designed with 39 springs, without combining separate networks).

Chapter TIME CHARTS,

SEQUENCE

AND GRAPHIC

7

DIAGRAMS,

DESCRIPTIONS

SIMPLIFIED OF RELAY

SCHEMATICS,

CIRCUITS

Anyone who has tried to understand from a schematic diagram the functioning of a relay circuit of even moderate size, recognizes that the task is complicated by the difficulty of following circuit paths through a large number of parallel and crossing lines. Furthermore, even if every circuit path is traced through, the dynamic action pattern of a circuit is not shown by the schematic diagram. To assist in understanding relay circuits, schematic diagrams are, therefore, supplemented by sequence diagrams, simplified schematics, and written descriptions. Of these, certain forms of sequence diagram and simplified schematic can be of considerable assistance in the actual designing of circuits. Use has already been made, of course, of simplified schematics and of switching algebra, which can be considered a simplified notation. However. neither of these gives any sense of time relationships, which are of primary importance in the sequential circuits to be discussed in the next chapter. In order to prepare for an understanding of sequential circuits. this chapter will discuss a variety of time charts or sequence diagrams which are useful for circuit design, analysis, and explanation; and, in addition, this chapter will present other forms of simplified circuit notation. In the design of sequential circuits, the actual time elapsing between successive events is usually less pertinent than the relative sequence in which events take place. In establishing relay control paths in sequential circuits, the chief concern is that a relay act before some and after other relays act. This is also true in the analysis of such switching circuits. Actual times, of course, must conform to requirements and must be consistent with apparatus capabilities. Also, actual time is often a major concern when a circuit is integrated with other circuits into a system. Time graphically are related two general scale while

charts and sequence diagrams are a means for indicating the actions taking place in a circuit, and how these actions to each other in time. The chief difference between these types is that time charts are drawn to an accurate time sequence diagrams are not. Both indicate the sequence in

- 131 -

The Design of Switching

132

Circuits

which events occur. Time charts are used in determining the actual work time of a series of circuit actions. Sequence diagrams are more often used in the design or analysis of switching circuits where a variable time scale permits expansion to show detail in a sequence of rapid actions.

(Cl

(Al

I-

-

I-=-

I-=-

n---0----------- (Kl

Fig. '1-1 A Sequentlal Circuit

7.1

SEQUENCE DIAGRAMS AND TABLES

Sequence can be shown in two ways. In one , the "state" type , a continuous record in chart or table form shows all pertinent input, output, and internal relay conditions as present or absent in every interval, and thus shows the combinations of relays operated at all times. The other, the "differential" type, shows only the changes of condition in their proper sequence. To determine, on the differential-type diagram, the state of a circuit in any interval requires a review of the circuit action in the preceding intervals. The state-type diagram, in turn, has the disadvantage that when the number of circuit elements is high, so much of the diagram is taken up by unchanging state indications that it becomes difficult to locate the changes and to follow their sequence. For this reason, the state diagram is usually used with small circuits and the differential diagram with large ones. Indication of circuit control means can be added to either type ·diagram by additional lines or symbols. Since any additional lines will tend to make the sequence less clear, whether or not they are used will depend on the particular application. Since no method of denoting sequence has been universally adopted, many variations in both diagrams and symbols are in use. Several representative methods will be described.

Sequence Diagrams

and Tables

133

State-Type Sequence Diagram. The first diagram to be considered is of the state type. The circuit used for illustration is shown on Fig. 7-1 and the corresponding sequence diagram is on Fig. 7-2. This diagram is divided into spaces of arbitrary length by vertical lines. Each vertical line represents the moment at which some action, such as the operation or release of a relay, takes place, with sequence progressing from left to right. The "state" of the circuit is indicated in the space between vertical lines. (In a relay circuit, the "state" is the combination of relays operated, and the action of a relay in operating or releasing changes the state.) Relays are assigned horizontal levels and are shown operated (not necessarily energized*) by a solid line on the horizontal level or are shown released by no line on that level t. The solid line runs through all spaces during which the relay is operated. Each vertical space can be numbered for reference. Note that a space must be left between any two actions which do not occur simultaneously. Input conditions, output conditions, keys, and other apparatus can be each assigned to a horizontal level, and their presence or operation indicated in the same manner as for a relay. To bring out the usefulness of such a sequence diagram, first a written description of the circuit in Fig. 7-1 is given as follows. (This circuit is used merely for example and is not shown as performing any useful function.) Closing key (K) applies ground to the winding of relay (A), operating it. Relay (A) closes the operating path (A+ C') which operates relay (B). Relay (B) locks on path (B + C'). Releasing key (K) releases relay (A). Now a path (A' + B) operates relay (C) and a second path (A' + B) holds relay (B). Relay (C) locks on path (B + C) and operates relay (A) over path (C + B). Relay (A) opens the locking path of relay (B) and holds relay (C) through path (A+ C). Relay (B), releasing, releases relay (A) which in turn releases relay (C). The sequence of operations when shown as on Fig. 7-2, although this diagram. Notice that interval relay (A); interval (3) the operate

described above is more apparent control paths are not indicated on (2) represents the operate time of time of relay (B); interval (5) the

• A relay is "energized" when an external source of potential sufficient to cause eventual operation is applied across Its winding. Some delay will always exist between this application and the building up of sufficient flux to move the armature. A relay is "operated" when make-contacts are closed and break-contacts open, and "released" when the converse Is true. Thus a relay can be energized and still be in a released position immediately prior to operating, or it can be de-energized and in an operated position immediately prior to releasing.

t This is sometimes

varied by designating each relay by a horizontal line. When the relay ls operated, the line is offset upwards by a slight amount, and is returned to the original level when the relay is released.

134

The Design of Switching 2

3

5

4

8

7

6

9

Circuits

10

I(

K A

B C 1

l

1

A

2

0

1

1

a

3

0

0

1

1

4

0

0

0

1

0

0

I

C

5

Fig. 7-2 State-Type

2

3

Sequence Diagram for Circuit of Fig. 7-1

4

5

6

7

a

I9

10

I(

0

1

1

0

0

8

0

0

0

9

0

6

1

7

1

10

0 0

A 8

C Fig. 7 -3 Modification of Fig. 7-2 to Show Control

Fig. 7-4 Sequence Table for Circuit of Fig. 7-1

Fig. 7-5 A Larger Sequential Circuit

release time of relay (A); and so forth. If controlling elements are to be identified, light arrows may be added as in Fig. 7-3. These show, for example, that the operation of (K) causes the operation of (A) which, in turn, causes operation of (B). This type of sequence diagram, without control arrows, will be used extensively in Chapter 8. Tabular Sequence Chart. A second type of presentation of sequence is one which uses the symbols introduced in Chapter 5 for the tabulation of relay combinations. Since a circuit with relays operating

Sequence Diagrams

and Tables

135 2

in sequence is progressing through various combinations of operated and unoperated relays, the sequence can be recorded by tabulation of the combinations in the order in which they occur. In these tables, time will run down the page, and each relay will be given a vertical column. A relay is shown operated by symbol O and unoperated by symbol 1. The action of Fig. 7-1 can be expressed by this method as shown in Fig. 7-4. Note that the numbered lines correspond to the numbered intervals of the diagram of Fig. 7 -2. The tabular sequence table will also be used in Chapter 8.

3

K

L M

C

A,B

N

02

0

0

F-

p Ml

Q

R

Fig. 7-6 Differential-Type Sequence Diagram of Circuit of Fig. 7-5

Differential-Type Sequence Diagram. The third method of presenting sequence is of the differential type showing only the changes in state. As mentioned earlier, this type of diagram becomes more necessary with large numbers of relays, since each relay is indicated and occupies space on the diagram only at the time it acts. Sequence progresses downward on this diagram, and actions are denoted by symbols. Relay operation is represented by "x" and release by "- ", accompanied by the designation of the relay involved. If an action is responsible for a subsequent action, the dependent action is indicated directly below the first and connected to it by a vertical line. When two parallel chains of action are both necessary for one action, that action is placed directly below one chain of action and connected to it by a vertical line. Then a line is drawn diagonally to this vertical line from the bottom of the other chain of action. A comparison of Fig. 7-5 and its sequence diagram of this type, Fig. 7-6, will demonstrate how these symbols are used. The coordinate system of letters and numbers included with this type of sequence diagram is for the purpose of making possible quick references. An application is shown at coordinate point o

.020

.o4o

.060

.oeo

.100

.120

.140

.160

.1110

I(

A

B C Flg. 7-7

State-Type

Time Chart for Clrcult of Flg. 7-1

.200

The Design of Switching

136 M-1 where relay (C) operating operate at Q-2. 7 .2

is a necessary action

for

Circuits

relay

(G) to

TIME CHARTS

Very often the length of time that a circuit and certain of its components will take to operate or the time that elapses between two of its output conditions. must be computed in order to determine how it will work in with other circuits and into the system of which it is a part. The values of the acting times of relays can usually be determined and the maximum, minimum, and average times may be computed for the whole or any part of the circuit operation if the corresponding numerical time values of the relays are added properly. If the action is at all complicated. however, it is much more convenient to draw a sequence diagram to a definite time scale. Thus, Fig. 7-2 could be redrawn to scale as in Fig. 7-7 where each vertical space column represents 0.010 second. Any desired time can be easily obtained by measuring it on the diagram. For example. the time between two successive operations of relay (A) is seen to be 0.116-0.20 = 0.096 second. A second method represents the acting time of a relay by segments of horizontal lines which are not assigned to a given horizontal level and which are shown only during the acting time. These line segments are identified by the relay designation, with some mark such as an underline to differentiate between operation and release. Vertical lines denote direct control; slanting lines with arrows. indirect control. A time chart by this method for Fig. 7-5 is shown on Fig. 7-8. The same type of differential representation may also be used with an arbitrary time scale.

0

.010

I

.020

-030

.040

.050

.060

.070

C

B A

G F

Fig. 7-8 Differential-Type

Time Chart Showing Control (or Circuit

or Fig.

7-5

Simplified 7 .3

Schematics

PARTIAL

and Graphic

Descriptions

137

SCHEMA TIC DRAWINGS OR "SHORTS"

In addition to means for presenting a clear picture of sequence of relay operations, there is often a need for showing relay control paths in a manner simpler than that of a complete schematic drawing of the circuit Usually. when only the contacts necessary for a given operation are shown, what had appeared as an involved path on the complete schematic diagram now appears as an obvious arrangement. This result. as mentioned in the opening paragraph of this chapter, may be achieved to a degree by the algebraic expression; however, where contact sequences. bridge-type circuits. and relay sequences are encountered, algebra does not always produce an adequate representation. Therefore resort is made to simplified diagrams called relay "shorts". The procedure used for preparing a relay "short" is to omit from the diagram all contacts not involved in the operation under consideration. to draw out the relay control paths in order, and to disassociate relay contacts and relay windings, with, perhaps, the exception of locking contacts. Instead of drawing each contact. an "x" may be used to represent a make-contact and "I" to represent a break-contact. Relay spring numbers may be included to identify particular B lAit )(::L _c--)~(---~1-._ (Cl -=contacts if the diagram is to be -48 VOLTS used with actual apparatus. Fig. 7-9 shows the operating path of Flg. 7-9 Short Schematic relay (C) of Fig. 7-1 drawn out (or a Relay from Fig. 7-1 in this manner. 7 .4

GRAPHIC DESCRIPTION

OF RELAY

CIRCUITS

A simple but complete representation of the action of a relay circuit can be of great aid to both the designer of the circuit and to others who, for various reasons, need to learn its operation. The graphical sequence diagrams discussed in the first part of this chapter are useful in presenting what circuit actions take place and when they occur, but some abbreviated circuit sketches or relay shorts of the type described in Section 7 .3 are needed to round out the presentation by showing how the actions take place. A substantially complete graphical representation of circuit operation by a combined sequence diagram and relay "short" is described in the following paragraphs. This graphic description, whose symbolic notations are shown on Fig. 7-10, makes use of a sequence diagram of the state type and is similar in this respect to the first diagram described in section 7 .1. As in that diagram, relays and other apparatus are assigned horizontal levels, with the state of any particular circuit component denoted by

The Design of Switching

138

I

SYMBOL

~

MEANING RELAY (Al ENERGIZED RELAY (Bl NOT ENERGIZED

A B

w

I

PATH Q.OSED

I I I

l£J

PATH OPENED

I

LQJ

w ~ ~

0

ACTIVE CONTACT

X

PASSIVE CONTACT I I I

J

t

et\

1

I I I

~

w

Circuits

DIRECT CONTROL PATH CLOSED TO ENERGIZE (Al, THEN OPENED

RELAY

SHUNT CONTROL PATH OPENED TO ENERGIZE (Bl, THEN CLOSED

RELAY

TYPICAL CIRCUIT PATH CLOSURE PASSIVE MAKE CONTACT ON RELAY (Al ACTIVE MAKE CONTACT ON RELAY (Bl NO CONTACT ON RELAY (Cl PATH CLOSURE ENERGIZES RELAY ID) BREAK CONTACT ON RELAY (El

17==

TYPICAL CIRCUIT PATH OPENING

A B

}RELAYS

NOT CONCERNED IN OPENING OF PATH

C

D--,

OPENING OF PATH DE-ENERGIZES RELAY ACTIVE BREAK CONTACT ON RELAY IE)

E---0---

CLOSURE OF LOCKING PATH FOR RELAY RELAY (Cl

ID)

(D)

BY

RELAY (B) CLOSING ITS OWN LOCKING PATH THROUGH RELAY (A)

AL----+-

OPERATION OF SLOW OPERATE RELAY (A)

RELEASE OF SLOW RELEASE

RELAY

(B)

ACTIVE PRELIMINARY MAKE CONTACT ON RELAY WITH ADDITIONAL ACTIVE BREAK AND MAKE CONTACTS

(Al

ACTIVE MAKE-BREAK TRANSFER CONTACT ON RELAY (B) WITH AN ADDITIONAL MAKE CONTACT

_9JB4++-

OJ

SYMBOL TO REFER TO NOTE I

Fig. 7-10 Symbols for Graphic Circuit Description

Graphic Description

of Relay Circuits

139

the absence or presence of a solid horizontal line at the corresponding level. Sequence progresses from left to right. (The orientation of the whole diagram with respect to horizontal and vertical may, of course, be interchanged.) To show circuit details, control path indications are included as vertical lines in the proper sequence locations on the diagram, whenever the paths are opened or closed. A solid vertical line is used to denote the closure of a path, and a dashed line to denote the opening of a path. The contacts making up the path are shown by symbols at the appropriate intersections of the horizontal relay levels and the vertical control path lines. The user of the diagram has available, therefore, information concerning both the over-all operation pattern and the particular causes and effects of every individual action. The number of symbols used has been kept as low as possible to facilitate both the construction and reading of the diagram. Fig. 7-10 lists all of the common symbols. The arrowhead is used to identify the control path line controlled circuit component. It is put on the vertical at the level of the component controlled by that path. Contact symbols differentiate between active and passive contacts. The active contact, whose symbol is a small circle, is the last one to close or the first to open a path, if a number of series contacts are involved in the path. On the other hand, the passive contacts, whose symbols are x's, are those which are necessary to complete the path, but which have been closed prior to the time at which the path is completed. A given contact may be a passive contact when a path is closed and may be the active contact when the path is opened. Whether or not a contact is a make type or break type and whether or not it is opened or closed can be determined by observing the state of its associated relay and the condition of the path of which it is a part. Thus, an active contact of a relay which is shown, after that relay has been released, as closing a control path is evidently a break contact.

Fig. 7- lOH shows a typical circuit path being closed by an active contact on relay (B) and passing through passive contacts on relays (A) and ( E). Note that the solid horizontal line showing energization of relay (D) starts immediately at the arrowhead. Fig. 7-101 shows the same path opened, but this figure does not have to show passive contacts since such contacts do not contribute to the path-openingoperation. Other symbols for slow relay operation, shunt paths, contact sequence, and the like, are explained in Fig. 7-10. Symbols for the active contacts of a relay are always shown adjacent to each other following the change of state of the relay. No inference is drawn concerning the sequence of action of any of the contacts unless specific symbols denoting contact sequence or relay speed are shown. To keep the number

The Design of Switching

140

~.__

Circuits

__________ __,

CR)

-=-

p

T

H

I

TO CIRCUIT •x•

Fig. 7-11 A Sequential Circuit

of symbols small, unusual circuit configurations bolically. but in the form of circuit notes.

are not covered

sym-

A difficulty with the state· type of sequence diagram, which has already been mentioned is that when the number of circuit elements is large, the diagram is cluttered up with unchanging state indications which conceal the significance of important sequences. To correct this situation, the graphic circuit description brackets those relays which remain relatively inactive for many intervals. The bracketed relays are then carried as a single state line for their inactive period, thus freeing space for the representation of other relay actions. When one or more elements within the bracketed group becomes involved in a circuit control path. it is indicated by the proper symbol on the group state

KEYS:K

9

9

I I

I I

RELAYS:A

9

B

EXTERNAL •x•------ll~---~---...&...--...L.-------11.....---CIRCUIT

[jJ

(I]

00

START SIGNAL TO CIRCUIT

rn

"X"

~ 2ND KEY CLOSURE TO CIRCUIT "X"

00 STARTS TIMING

FOR RELEASE OF IA)

Fig. 7-12 A Graphical Representation

of the Circuit of Fig. 7-11

Graphic Descriptions

of Relay Circuits

141

line. External circuits are carried in a similar manner by single state lines. The method can be extended to show the interaction among minor or major components of a system instead of apparatus elements. To conclude. an example of the graphic circuit description is attached as Fig. 7-12 representing the operation of the circuit of Fig. 7-11. The circuit itself is intended only to illustrate the graphical notation and performs no obvious function.

Chapter 8 DESIGN OF SEQUENTIAL RELAY CIRCUITS In a combinational circuit, each input combination invariably produces the same output combination every time it occurs. In a sequential circuit, however, the output conditions are determined jointly by the sequence in which input signals occur as well as by the combinations of input signals. The action of a sequential relay circuit depends upon arrangements of relays and circuit paths which recognize certain input combinations, retain a memory of the fact that these combinations have occurred, and use this memory to influence later circuit actions. Input signals to a sequential circuit take the same forms as those for combinational circuits. In the following discussions it is assumed that there is one input lead for each input signal and that the leads connect directly to relays in the circuit. These relays, which will be called "primary" relays, perform a function similar to that performed by relays in a combinational circuit, in that they detect input signals and may establish a wide variety of circuit paths dependent on the input combinations. A sequential circuit also contains one or more relays called "secondary'' relays which provide a record of significant actions the circuit has taken, and thus guide future circuit operations. Secondary relays are controlled within the sequential circuit by local circuit paths which may employ contacts on both the primary and secondary relays. The circuit for each secondary relay is or the locking type where the relay participates in its own control. In many cases, the circuit can be manipulated so that one or more primary relays require but a single make contact and therefore can be eliminated from the circuit. In the extreme situation, no relays identifiable as primary relays appear at all in the circuit. Sequential circuits can be illustrated by a simple example. Suppose that a signaling system between points A and B has the following requirements. Grounding a lead a by means of a key at position A is to cause a lamp to light at position B. U a lead b is then grounded by means of a key at B, the lamp Is to be extinguished. The lamp must remain extinguished U the key at B is restored to normal before the key at A Is restored.

If the input leads control primary relays (A) and (B), the operating sequence can be indicated as in Fig. 8-1. During interval 2, when

- 142 -

Design of Sequential

Relay Circuits

2 PRIMARY RELAYS

3

143

4

2

A PRIMARY RELAYS

B

SECONDARY RELAY

LAMP

3

4

A

B

X

LAMP

Fig. 8-1 A Sequence of Primary Relay Operations

Fig. 8-2 Sequence of Action of a Secondary Relay

the circuit to the lamp must be closed, relay (A) is operated and (B) is released and the circuit path (A + B') will be closed. However, this same input combination exists in interval 4 when the lamp must be extinguished, and no arrangement of contacts on relays (A) and (B) alone will permit different functions to be performed in intervals 2 and 4. A solution is to provide a secondary relay which will be in an operated state during one of the intervals in question and in a released state during the other. Considering all relays, both primary and secondary, the combinations then will be different during intervals 2 and 4, and a circuit path for the lamp can be developed. A suitable method of controlling the secondary relay in this example is to let it operate during interval 3 when (B) operates, and lock to a make on (A) so that it will hold through interval 4 and release after (A) releases. This sequence is shown in Fig. 8-2 where the relay is designated (X). The line in the chart for relay (X) indicates the operation of the contacts of this relay. Although (X) starts to operate at the beginning of interval 3 when (B) first operates, there is a short operate time

a

-!= o

J (A)

(Bl

-P ___._ b

-r

Fig. 8-3 Circuit

for Sequence of •Fig. 8-2

The Design of Switching Circuits

144

interval before the contacts act to open or close circuit paths. In a similar manner, the relay contacts remain operated for a short time after interval 4 while the relay is releasing. A circuit to fulfill the requirements can be easily constructed and is shown in Fig. 8 -3. A break contact on relay (X) prevents the lamp from lighting during interval 4. Thus, a secondary relay controlled by the primary relays to operate in a prescribed sequence provides control means which cannot be achieved by the primary relays alone. 2

3

4

A B

RELAY IX) ENERGIZED RELAY IX) OPERATED LAMP

Flg. 8-4 Energlzatlon of Secondary Relay (X)

8.1

CONTROL OF SECONDARY RELAYS

In designing a sequential circuit, the number of secondary relays required must be determined first; then a suitable operating sequence for these relays must be chosen; and finally, control paths to cause the relays to act in this sequence must be developed. Although the design of control paths is usually the last stage in the development of a circuit, it will be discussed first, since a knowledge of control paths is helpful in developing satisfactory operating sequences. Assume, therefore, that a satisfactory sequence diagram showing the action of all primary and secondary relays has been developed for a circuit problem. By examining the diagram, the combination of relays operated at any given time can be determined. To develop a circuit path which closes when some circuit action occurs, and remains closed until some later action takes place, it is necessary only to observe what relay combinations exist during the time the circuit path is to be closed, and to proceed by combinational methods. For instance, the circuit for controlling a secondary relay can be developed by determining the intervals during which its winding must be energized, and developing a circuit path closed in these intervals. As will be seen, the circuit paths can be developed by methods comparable to the formal approach discussed in Chapter 6, or by the method of inspection.

Design of Sequential

Relay Circuits

145

0

b

I Fig. 8-5

Control of Relay (X) by (A+ BX)

Consider, as a first example, the control of relay (X) in Fig. 8-2, which shows the sequence of action of contacts on relays (A), (B), and (X). In Fig. 8 -4 a dashed line has been added through intervals 3 and 4 to indicate when a circuit path must be closed to energize the winding of relay (X). This does not correspond to the line on the chart showing the action of the contacts of (X), since the acting time of relay (X) subdivides interval 3 into two parts. For the same reason there are two parts to the interval following 4 (in which the circuit returns to the normal condition shown in interval 1). The operate and release times of secondary relays are often of considerable importance, and their existence should always be recognized in the design of sequential circuits. From the three relay combinations existing and 4, a circuit for (X) can be developed as follows:

during

intervals

3

f(X) = (A + B + X')(A + B + X)(A + B' + X)

The circuit

= (A + B) (A + X)

(8-1)

=A+

(8-2)

BX

is shown schematically

in Fig. 8-5.

As with combinational circuits, invalid relay combinations may be included to achieve simplification. In this case relay combinations (A' + B + X') and (A' + B + X) never occur in the operating sequence. These combine to give (A' + B) which may be combined with equation (8-1) as follows: f(X) = (A + B)(A + X)(A' + B)

= B(A + X) This path was shown schematically

(8-3) in Fig. 8-3.

The Design of Switching

146

Circuits

(LAMP)

a

b

Fig. 8-6 Elimination of a Primary Relay

Algebraically, equation (8-3) is not equivalent to equation (8-2). but either contact arrangement will provide the proper control path for relay (X). When such alternative choices occur, the ease of combining with other paths usually determines which will be used. In this example the operating path for the lamp is: f(L) = A + B' + X' This can be combined follows: f(X) =A+

f(L) =

with the circuit

for (X) in the form

(8-2)

as

[BX B' + X'

Since relay (A) contains only a single make contact, it is unnecessary to include that relay in the circuit. The resulting simplified circuit is shown in Fig. 8-6. The method used above, that of employing the combinations in HOLD PATH (X l all intervals during which the conQ trolled relay is energized, is genOPERA~TE PATH eral and could be used to design the ½P I•~ control path of any secondary relay. However, it becomes unwieldy Flg. 8-7 where many relays and combinaBasic Secondary Relay Control Paths tions are involved, and a more direct approach is desirable. The circuit of a secondary relay is basically of the locking type shown in Fig. 8-7, which may be written:

--:i

f(X) = P(Q + X)

(8-4)

The network P is the "operate path" which must close at the time the relay (X) is required to operate. The network Q, which is connected in series with the locking contact X, is the "hold path" which must close during or before the closure of P and must remain closed until the relay is required to release. The design work may be reduced by developing each of these paths separately. The circuit then can often be

Design of Sequential

Relay Circuits

147

simplified by combining elements previous example, equation (8-1):

in the operate and hold paths. In the

f(X) = (A + B)(A + X), the operating path is (A + B) and the holding path in series with the locking contact X, is A. These were combined in equation (8-2): f(X) =A+

BX.

If there are no restrictions on the number and type of contacts available on the relays of the circuit. the control path for a secondary relay can always be developed in the general form given in equation (8-4). Cases arise. however, where shunt control or other special circuit arrangements are used because of practical restrictions.

It can be observed that the operate path P and the hold path Q in the general form f(X) = P(Q + X), will contain neither make nor break contacts on relay (X). The circuit combination which causes relay (X) to operate must come into existence before (X) operates and will be followed by a combination which is identical except that (X) is operated. Thus the operate path P obviously does not require contacts on (X). Since the hold path Q is connected in series with a make on (X), any contacts on (X) contained in Q can be eliminated by algebraic methods. Since operate and hold networks, P and Q, tfo not then contain contacts on the relay being controlled, these networks may be designed by considering the combinations of all relays of the circuit except the one being controlled. The operate path of a secondary relay must be closed in the interval in which the relay is to move to the operated position. However, this path can often be simplified by permitting it to be closed for other combinations of the circuit relays than the one existing at this time. It may be closed during any of the intervals in which the relay will remain operated. and may also be closed for any of the invalid combinations which do not occur in the normal sequence of actions. It must not be closed for any combination in the sequence in which the relay is expected to remain unoperated. Although the circuit may be designed by symbolic methods, it can often be determined from a direct inspection of the operating sequence. The simplest statement of the position of other relays which identifies the point in the sequence in which the relay acts will indicate- the form of the circuit. The hold path of a secondary relay must be closed in every interval during which the relay is energized excepting, possibly, those in which the operate path is closed. Since the hold path is in series with a make contact on the relay, it may be closed for any combination of the circuit relays except the one which occurs at the time the relay must release. One design method is to determine those relay combinations

148

The Design of Switching

Circuits

for which the relay must hold operated, and to construct a network closed for these combinations. A second method is to develop a circuit path which opens only at the time the relay must release and is closed at all other times. The combination which exists during the interval in which the relay releases and which must cause the holding path or the relay to open is called the "release combination". A circuit network open only for this combination (the negative or the circuit closed for this combination) can be used as the holding network. Holding circuits developed by these two methods will not necessarily be the same• although both fulfill essential requirements. The first is closed for the minimum number or combinations required to hold the relay operated, while the second is closed for all except the release combination. The simplest circuit often lies between these two limits. Circuits developed by the first method often can be simplified by including ce!'tain of the remaining combinations in the conditions for closing the path. Invalid combinations not in the normal sequence as well as combinations existing while the relay is released may be used. The determination of a hold path by either or these two methods may not give the simplest path, since if there are many combinations available for simplifying, it is difficult to pick the proper combinations for the simplest result. For this reason, a completely different third approach based on inspection is often taken. After determining the relay's operating path, the sequence diagram is examined for a contact or several contacts in series which will serve to hold the relay from the interval in which the operate path is opened to the interval in which the relay is to release. If none can be found which will serve for the complete holding time, a path may be picked which covers at least a few intervals, at the end of which another path is selected to take over, and so on upto the release interval. Caution must be observed that each hold combination determined in this manner will be open when the relay is to release. The parallel connection of these paths will give the hold path, usually in a simple form. When a secondary relay operates two or more times in the sequence or circuit actions, the preceding methods must be modified. The operating network must close each time the relay is required to operate. It can be designed by developing a circuit path for each of the several operating intervals, and then placing the combined paths in parallel. The paths should be carefully checked to see that they do not close at times when the relay must remain unoperated. Similarly, holding paths can be developed for each of the periods during which the relay must hold operated, and these paths placed in parallel in the locking • It can be shown that when all combinations of all clrcult relays, both prlmary and secondary, occur ln the clrcult action sequence, clrcults obtained by both methods are ldentical.

Design of Sequential

149

Relay Circuits

circuit. Each holding path is required to be open not only at the end of the period during which it serves to hold the relay but also must be open during each of the other intervals in which the relay releases. If a relay is assumed to operate and release three times in a sequence, it will have three operating paths, P 1, P 2 , and P 3 , and three holding paths, Q1, Q2, and Q 3 . The combined circuit will be: f(X) = P 1 P2P3 (Q 1 Q 2 Q3 + X) The holding path may be developed by a second method. A satisfactory circuit will be one which is closed for all combinations except the ones during which the relay releases. This circuit can be developed by constructing a network for each release condition which is open only when this condition occurs, and connecting the networks for the several release conditions in series. If the release networks for the relay which acts three times are designated R 1 , R 2 , and R 3 the resulting circuit is: f(X) = P1P2P3(R1

+ R2 + R3 + X)

From the above discussion it is seen that a circuit to fulfill the control requirements of a secondary relay can be designed by considering only those intervals during which the relay operates or releases. be the simplest one. The resulting circuit, however, will not necessarily

2

A PRIMARY RELAYS

B

3

4

5

7

6

8

9

10

-

----

C

SECONDARY RELAY

Fig. 8-8

X

Sequence with Two Operations of a Secondary Relay

Fig. 8-8 shows a sequence ondary relay (X), once in interval sequence: Operate

Path=

(A+ [

requiring two operations of a sec2 and again in interval 7. For this

B' + C')(A'

first operation

J[

The path may be simplified (A+ B + C) as follows:

+ B + C)

second operation

J

by using

the

invalid

•

combination

The Design of Switching

150

Circuits

Operate Path= (A+ B' + C')(A' + B + C)(A + B + C) = (A + B' + C')(B + C) The hold path in intervals Hold Path=

3, 4, and 8 is determined

as follows:

[ (A' + B' + C')(A' + B + C')] (A' + B' + C) secon_d ] [ first operation operation

J

=A'+

[

B'C'

Then the complete circuit

is:

f(X) = (B + C)(A + B' + C')(X + A' + B'C') = (B + C)(A + B' + C')(X + A')

If the hold path is determined by the second, or release method (intervals 5 and 9), then:

combination,

Hold Path = (A + B + C')' + (A + B' + C)' = A'B'C + A'BC' = A'(B'C

+ BC')

= A'(B' + C')(B + C) And

f(X) =(A+

B' + C')(B + C) [X + A'(B'

+ C')(B + C)J

=(A+

B' + C')(B + C) [X + A'(B'

+ C')]

= (B + C)(A + B' + C')(X + A') Both methods give identical

results

in this example.

0

(C)

b C

(X)

..r:+1 •

Fig. 8-9 Clrcult for Fig. 8-8

Design of Sequential

151

Relay Circuits

The circuit must now be examined for operating hazards due to races or contact stagger sequences. A hazard due to a temporary break in the energization of relay (X) may occur on the release of relay (A) between intervals 2 and 3 when the path transfers from (A+ B' + C') to (A' + X). This can be overcome by using a continuity-transfer for the A-A' pair. The circuit may also be rearranged so that a break-make transfer is satisfactory by using the following theorem of switching algebra:

(X + Y)(X' + Z) = XZ + X'Y f(X) = ( AX + A' (B' + C')

The control

path then is:

The circuit

is shown in Fig. 8-9.

J

(B + C)

Circuits for this example can also be constructed directly from an inspection and analysis of the operating sequence. From the sequence diagram, Fig. 8-8, it is seen that (X) operates when (A) operates in interval 2, but does not operate when (A) re-operates in intervals 5 and 9. The condition of both (B) and (C) released is sufficient to distinguish 5 and 9, since (B) is operated during interval interval 2 from intervals 5 and (C) is operated during interval 9. Therefore (X} must operate when (A) operates and both (B) and (C) are released, i.e. (A+ B' + C'}. Relay (X) must also operate in interval 7, which may be described as the interval of the sequence in which both (B) and (C} are operated, i.e. (B + C). For the holding path, it is seen that the relay could hold to a break on (A) during intervals 3 and 4, and release during interval 5 when (A) operates. However, the hazard of the relay releasing due to a momentary open during the change from interval 2 to interval 3 is immediately recognized. One solution would be to include in the locking path a circuit which is closed continuously during intervals 2 and 3. From the diagram it is seen that the condition "(B) and (C} released" exists continuously during intervals 1, 2, and 3-, and will not interfere with the release of the relay during intervals 5 and 9. The path (B' + C') therefore can be included in the locking path. The single break on (A} is satisfactory to hold the relay in interval 8 during its second operation, since this contact is closed in interval 6 before the relay operates and remains closed through intervals 7 and 8. From the above analysis the following circuit can be written:

f(X} = (A+ B' + C') [X + A'(B' + C'}] (B + C} This can be simplified

as follows:

f(X} = [A+ A'(B' + C'}] [X + A'(B' + C'}] (B + C} = [AX+

A'(B' + C'}] (B + C}

This is the same as the circuit methods.

in Fig. 8-9 which was obtained by formal

152

The Design of Switching Circuits 2

3

5

4

6

7

8

9

10

A PRIMARY

RELAYS

B C

-

----

X SECONDARY RELAYS y

Fig. 8-10 Sequence with Two Secondary Relays

A

B C

D

1

2 3 4 5

6 7 8 9

0

1

1

0

0

1

0 0 0

1 0 0 0 0 0 0 1 0 0 0 1 1 0

10

0

11

0 0 0

0

12

0 0 1 1 0 1 1 1 0 1 0

0

13 14 15

16

0

Table of Sequential Requirements

1 0

0 0 0

Circuits considered so far have contained only one secondary relay. Design approach is not materially altered when several secondary relays are involved. The control of each relay is considered separately, assuming that other secondary relays operate in the desired sequence even though their control paths may not yet have been designed. With all paths independently completed, the circuit will act in the sequence indicated. For example, take the sequence of Fig. 8-10. Since the condition of (C) operated and (Y) released in interval 4 is unique, the path (C + Y') can be used for the operation of secondary relay (X). The relay can be locked during intervals 5, 6, and 7 to a break on (C) (C) giving the circuit: f(X)

==(C

+ Y') (X + C')

Rearranging on (C), f(X)

to avoid a continuity-transfer

ex+ c•y•

==

Similarly for the control of relay (Y), the path (B' + X) which is closed during interval 6 may be used for the operating path, and the relay can be held through interval 10 by locking to a make on (A). The circuit is: f(Y)

== (B'

+ X)(A + Y)

Design of Sequential Relay Circuits

153

The complete circuit for relays (X) and (Y) is shown in Fig. 8-11. Note that relay (A) can be eliminated without affecting the operation of the circuit.

Flg. 8-11 Circuit for Flg. 8-10.

Much of the combination and simplification of the previous examples could have been done by the tabular method. If the number of relays is not too large, the tabular method of denoting sequence gives a direct approach to the subsequent contact manipulation. Operate, hold, and release combinations can be selected from the table by observing where the secondary relay to be controlled changes its state. The tabular method will be used for a final secondary relay control example in which all the relays are internally controlled and which will run through its sequence cycle continuously as long as power is provided. This particular sequence, shown in the table on the opposite page, uses all combinations of four relays. The holding paths will be determined from the combinations which cause the relay to release. Relay (D) operates and releases once in the cycle. It is energized in interval 7 by combination (A+ B' + C), and released in interval 15 by combination (A' + B + C'). The holding path, therefore, is the negative of this latter path.

The Design of Switching Circuits

154

Flg. 8-12 Recycling Clrcult on Four Relays

f(D) = (A+ B' + C) [D + (A' + B + C')'] =(A+

B' + C)(D + AB'C)

The circuit for relay (A) is determined from intervals the operating path, 4 and 11 for the holding path,

1 and 6 for

f(A) = (B' + C' + D')(B' + C + D')(A + B'C'D + B'C'D')

= (B' + D')(A + B'C')

The circuit for relay (B) is determined from intervals 14 for the operating path; 5, 12, and 16 for the holding path,

2, 9, and

f(B) =(A+ C' + D')(A + C' + D)(A' + C' + D) • (B + AC'D + AC'D' + ACD)

= (C' + AD)[B + A(C' + Dtl = C' B + AC I + AD = C'(A + B) + AD

= (C' + AD)(A + B)

(Notice that the holding path for the third operation of relay (B) is elim·inated by algebra since the relay releases when its operate path opens.) The circuit for relay (C) is determined from intervals for the operating path; and from 8 and 13 for the holding path,

3 and 10

Design of Sequential

Relay Circuits

f(C) = (A+ B + D')(A

155

+ B + D)(C + A'BD'

+ ABO')

= (A + B)(C + BD')

The combined circuit for all four relay paths is shown in Fig. 8-12. There are situations in this circuit where a relay acts through For example, the contacts of another relay which it in turn controls. relay (B), which operates through path (A + C') in interval 3, immediately operates relay (C) to open its own operate path. This requires that contact stagger on relay (B) be less than the operate time of relay (C). If this requirement can not be met, a preliminary make contact must be used for locking (B) and other relays in similar situations. 8.2

EFFECT

OF RACE CONDITIONS

ON INTERNAL

CONTROL

In many sequential circuits, one relay by its operation may affect several relays at the same time. This causes a "race" condition, and the sequence in which the relays act is uncertain. These conditions must be examined to determine whether some relay combination which may exist momentarily may cause adverse reactions. In cases where the exact sequence of action of several relays is unimportant, simultaneous operation may permit simpler relay control paths and may result in the circuit performing its functions in a shorter time.

A sequence diagram with several race conditions is shown in Fig. 8-13. In interval 3, when both (X) and (Y) operate due to the release of (K), there will al ways be a slight difference between the times when (X) and (Y) reach the operated position. Thus either combination (K' + W + X + Y' + Z') or

(K' + W + X' + Y + Z')

2

PRIMARY I( RELAY

3

4

- -

5

6

7

I

- -

9

w SECONDARY RELAYS

X y

z Fig. 8-13

Sequence wlth "Simultaneous"

Secondary Relay Actions

The Design of Switching

156

Circuits

will exist for a short time in interval 3. The effect of these two combinations must be considered when developing circuit paths, and neither can be treated as an invalid combination. If the sequence were modified to operate relays (X) and (Y) in a prescribed order, one of these two combinations would be specified and the other then could not occur. For example, (Y) could be operated through a make contact on (X) and combination (K' + W + X' + Y + Z') would never occur. The same sequence than relay could be achieved by making relay (Y) slower in operating and (Z) operating. (X). In interval 5 there is a race of (X) releasing Other races occur in intervals 7 and 9.

8.3

PROVIDING OUTPUT CONDITIONS IN SEQUENTIAL

CIRCUITS

The output conditions which a sequential circuit establishes may last for one interval only, or extend over several intervals, or appear several times. In each interval the problem of establishing the proper is output conditions is of the combinational type. The only difference that when an output condition extends over several intervals, care must be taken that the condition is maintained continuously. This problem is the same as that of maintaining continuous hold paths for secondary relays, and may require contact manipulation or use of continuitytransfers.

8.4

DEVELOPING

AN OPERATING SEQUENCE

From the requirements of practical design problems, it is usually possible to determine the sequence in which input conditions occur and the corresponding sequence in which output conditions must be produced. rrom these sequences the designer must determine when secondary relays are required, and choose suitable intervals in the sequence for the operation and release of these relays. This work must precede the design of the actual relay control paths and is often the most difficult phase of circuit design. Use of a diagram in developing the sequence will be found helpful. The sequence will start with a "normal" condition and proceed through the sequence until all actions are completed and the circuit has returned again to the normal condition. This is a circuit ''cycle". The normal condition is usually the one which exists when the circuit is idle. In most cases this will be the condition in which no input signals are present and all of the circuit relays, including secondary relays which may be added during the design, are unoperated. There are, of course, exceptions. There are several types of operating sequences. The simplest is where the inputs and outputs al ways occur in a definite sequence. The examples which have previously been given in this chapter are of this

Design of Sequential

Relay Circuits

157

type. Another type of sequence is where the inputs, starting from a normal condition, may occur in several alternative sequences, and the output is influenced by the particular sequence in effect. For- thtse circuits, the sequence diagram should show all of the possible alternative cycles in some arbitrary order separated by normal intervals. Although the diagram will appear to state a definite order in which alternative events occur, the completed circuit will follow any one of the alternative cycles from the normal state whenever the input conditions are applied in the appropriate sequence. The normal intervals must all be kept alike in both primary and secondary relay combinations, so that no cycle is affected by the previous cycle through which the circuit has passed. The following

example

illustrates

an alternative

sequence:

An intercommunication system is installed between two positions, A and B. Each position is provided with signaling equipment consisting of a key of the mechanically locking type, an answer lamp (ANS-), and a disconnect lamp (DIS-). Operation of the key at one position will cause to light. This lamp is extinthe (ANS-) lamp at the other position guished when the key at the called position is operated. If either key is then restored to normal, the (DIS-) lamp at the other position will light and remain lighted until that key is restored to normal. A relay circuit for the control of the signal lamps is required.

Four alternative sequences of input signals are included in these requirements and are shown in Fig. 8-14. In each case the output depends upon the input sequence. Intervals 1, 5, 9, 13, and 17 are all normal with no input leads grounded.

PRIMARY

-

A

RELAYS B

ANS-A ANS-B OUTPUTS DIS-A DIS-B

-

-

-

-

* NORMALINTERVALS Fig. 8-14 Sequence with Alternative

Cycles

-

The Design of Switching

158 2

•

INPUT

OUTPUT

3

4

Circuits

5

ST

TONE

Fig. 8-15 Sequence Requiring Internally Generated Intervals

In analyzing the requirements of a sequential circuit, consideration must be given to all conceivable sequences. Of these a large number may usually be discarded as being known never to occur. In many cases, however, the statement of requirements may not mention some of the possible cycles. For example, in the previous statement, no mention was made of what action the circuit was to take if a key should be restored to normal without waiting for an answer. In practice, a designer would probably have to verify that a sequence such as this never occurred, or if it did occur, ascertain what action should be taken. In this case, it can be assumed that it may occur, and, if it does, the circuit should restore to normal after the restoring of the key. A circuit to operate in this sequence will be designed later in this chapter. The number of alternative cycles may easily rise too high to allow the complete representation of each cycle on a practical sequence diagram. For example, in automatic telephone systems all circuits involved in setting up a call from one subscriber to another must be capable at many or all stages of their operation to return to normal should the call be abandoned by the originating subscriber. This sort of general requirement involving a large number of alternative cycles is easily understood and is most usable in its simple verbal form rather than in a long series of repetitious sequence diagrams. In some situations, signals may occur on a group of similar input leads in a more or less random sequence, and the circuit is required to establish output conditions dependent upon the relative sequence in which the signals occur. An example is where several independent signal sources are associated with a single common device which can perform some function for the individual sources on a one-at-a-time basis. When one source obtains the service of the common device, all others must be made to wait until service to the first is completed. Waiting requests are then served in a prescribed order; It would be impractical in these cases to investigate every possible sequence in which the inputs may occur. Instead it is usually possible to proceed without a sequence diagram by determining a general plan for assigning a preference order to each individual input according to its physical

Design of Sequential

Relay Circuits

159

position relative to others which make simultaneous or prior requests. The circuit design then can usually be carried out by the methods which have been described for positional networks. Lockout circuits, discussed in a later chapter, are an example of circuits of this type. Another general type of circuit produces a sequence of output signals entirely by internal control means. In these circuits, an input from external sources merely starts the action, and subsequent intervals of the sequence are determined by elements within the circuit and do not correspond to changes of input signals. An example is a tonesignaling system having the input and output sequence requirements shown in Fig. 8-15, where ground applied to an input "ST" lead is to cause two spurts of tone to be applied to an output signal channel. The problem here is to choose relays and to determine an acting sequence in which the operate and release times of the relays will provide the proper time intervals. In some cases other devices, such as electronic tubes, will be used to generate the required time intervals.•

8.5

PLANNING SECONDARY RELAY ACTIONS

After the requirements of a sequential circuit have been determined, the next step in design is to plan the sequence in which secondary relays shall operate. The preliminary diagram of the sequence of primary relay operations and of the output signals serves as a convenient framework to which the sequence of secondary relay operations can be added. In order

to deliver

a distinctive output signal during a particular sequence, a circuit must provide a combination of operated relays which exists during this period and does not exist when the output signal is undesired. If the primary relays alone provide suitable combinations, secondary relays are not required. However, if there are intervals in the sequence which have identical input combinations but require different output combinations, secondary relays are necessary. These relays must be inserted in the operating cycle so that their combinations, taken in conjunction with the original combinations of the primary relays, provide unique relay combinations which can establish the desired output paths. The need for secondary relays, then, is recognized when an input combination appears more than once in a sequence and is required to produce different outputs on each appearance. As a minimum, the number of secondary relays must be sufficient to provide a distinct combination for each identical input interval which must be recognized. That is, one secondary relay can distinguish

period in an operating

• Speclflc technlques for developlng sultable sequences for a group of relays cussed ln the appendix to thls chapter, section 8.7.

are dls-

The Design of Switching Circuits

160 2

3

5

4

6

7

8

9

II

10

A PRIMARY RELAYS B

C

SECONDARY X RELAY

OUTPUT

I

I

M

Fig. 8-16 Sequence with Several Combinations

Available for Control

between no more than two identical input intervals, two secondary relays between no more than four identical input intervals, and so on. After recognizing that secondary relays are required, and determining which input intervals are identical and must be identified, other intervals in the cycle must be selected for the operation and release of the secondary relays. These operate and release intervals must have certain characteristics of uniqueness; otherwise it will be impossible to design circuit paths which will control secondary relays to act as desired. Clues to the selection of appropriate operate and release intervals for a secondary relay are furnished by the methods of designing operate and hold paths, discussed earlier in this chapter. In order to determine the validity of proposed operate and release intervals, it is sufficient to check only the corresponding operate and release combinations established by the states of all relays except the one being controlled. These combination·s must meet the following requirements: 1. The operate combination must not exist during any interval in which the relay must be released, but may be repeated during any interval in which the relay is operated. 2. The release combination must not exist during any interval in which the relay is operated, unless covered by an operate combination, but may be repeated during any interval in which the relay is released. Sequences which meet these requirements will produce workable circuits. In some cases which appear to require only a single secondary relay, the sequence may not contain suitable intervals for the control of this relay. It then becomes necessary to provide an additional secondary relay in order to establish combinations for the control of the first

Design of Sequential

Relay Circuits

161

relay. Often each secondary relay must participate in the control of the other so that a satisfactory sequence for either relay cannot be developed without considering the presence of both relays. Regardless of the sequence of occurrence or number of repetitions of combinations of input signals, it is always possible to identify a particular appearance of an input combination by providing enough secondary relays. Some of the problems encountered in selecting secondary relay sequences may be illustrated by the following examples. Fig. 8-16 shows a cycle with two input combinations each of which is repeated: (A+ B' + C') in intervals 2 and 8, and (A + B + C') in intervals 3 and 7. An output is to be established only in intervals 7 and 8. One secondary relay, (X), released through intervals 2 and 3 and operated through intervals 7 and 8 will serve to differentiate between the similar input combinations. Several combinations are available for the operation and release of this secondary relay; it may operate in intervals 4, 5, or 6 and release in intervals 9, 10, or 11 (normal). The governing factor in picking the secondary relay control intervals will be the simplicity of the control path. By inspection, if a single make contact on relay (C) is chosen to operate the secondary relay, it will operate the relay in interval 5 and hold it in intervals 6, 9, and 10. To hold it during intervals 7 and 8 requires only a locking circuit through a make contact on relay (A), which is effective through intervals 6, 7, 8, and 9. Since this overlaps with the operating path in intervals 6 and 9, continuity of the circuit is assured. The following simple control path is thus obtained: f(X) = C(A + X)

The output circuit f(OUT)

::1

which is closed during intervals

7 and 8 is:

A + C' + X

A sequence vals is restricted

in which the choice of secondary relay control interis shown in Fig. 8-17. Here, a secondary relay is 2

3

4

5

6

7

8

9

10

II

12

A PRIMARY RELAYS

8 C

OUTPUT

M

Fig. 8-17

Sequence with Two Combinations Available for Control

13

The Design of Switching Circuits

162

needed to differentiate interval 12 from interval 6 in order to establish the output condition in interval 12. As a first attempt, let the relay be in the operated state during interval 6 and in the released state during interval 12. It is evident that if the secondary relay is operated in intervals 2, 3, 4, or 5 and held through 6, it will also operate in one of intervals 8 through 11 and hold through 12, since the combinations of inputs from 2 through 6 are repeated in intervals 8 through 12. Similarly, if the relay is to be in the operated state during interval 12 and in the !eleased state during interval 6, it cannot be operated in any of intervals 8, 9, 10, or 11. This leaves only the combinations of interval 1 (same as 13) and of interval 7 available as control combinations. Assuming that it is undesirable to have the secondary relay operated throughout the normal interval, it must be operated in interval 7 and released in interval 13 as shown in Fig. 8-18. By inspection, the circuit is: f(X) =(A+ B + C')(X + AB) The output circuit is: f(OUT) = A' + B + C' + X or A' + C' + X 2

3

4

5

6

1

8

9

10

12

II

13

A

PRIMARY B RELAYS C

SECONDARYX RELAY

OUTPUT M

. .

. •

-

Fig. 8-18 Sequence of Fig. 8-17 with Secondary Relay Added

The sequence of Fig. 8-19 offers more difficulty in the search for intervals containing suitable operate combinations. In order to provide an output path which is closed in interval 13, but not in interval 7 where the primary relay combination is the same, first assu"me that a secondary relay will be in the operated state during interval 7. It is found by inspection that the sequence of relay combinations from interval 2 to 7 is the same as the sequence from interval 8 to 13. Therefore if a secondary relay is controlled by relays (A), (B), and (C) to operate in one of the intervals 2 to 6 and to hold through interval 7, it will repeat a similar action in the later part of the cycle and hold through interval

Design of Sequential Relay Circuits 2

3

4

s

6

163 7

8

9

10

II

12

13

14

15

A PRIMARY B

RELAYS

C

OUTPUT

-

M

Fig. 8-19 A More Difficult Sequence

13. Intervals 14 and 15 (or 1, the normal interval) are the only unique intervals in the cycle; and these, obviously, cannot be used to control a relay to be in the operated state in only one of the intervals 7 and 13. It is necessary, therefore, to provide an additional secondary relay which will make one of the intervals between 7 and 13 dissimilar from the corresponding one preceding interval 7. Once this is realized, there is a wide choice of intervals in which to operate and release the two secondary relays. One such choice, not necessarily the simplest, is shown in Fig. 8-20. Designating the secondary relays (X) and (Y) in the order of their operation, the circuit paths are: f(X) = (A + C)(AB + X) f(Y} = (X + C')(C + Y)

=ex+

C'Y The output circuit,

closed during interval 13, is:

f(OUT) = A + B' + C + Y I

2

3

4

5

6

7

8

9

10

II

12

13

14

A PRIMARY

RELAYS

B C

..

.

~

SECONDARY RELAYS

;~

-

-

OUTPUT

Fig. 8-20 A Sequence of Secondary Relays for Fig. 8-19

IS

The Design of Switching Circuits

164 2 PRIMARY RELAY

*SECONDARY RELAYS

4

5

-{ xz

-

-

OUTPUT TONE

*

3

ST

BOTH RELAYS X ANO Z MUST BE SLOW ACTING TYPE

Fig. 8-21 One Sequence of Secondary Relays for Fig. 8-15

The requirements of Fig. 8-15 present a different problem in sequence planning. Here, the acting times of secondary relays must generate intervals 3, 4, and 5. No secondary relay combination may be repeated, because then the circuit would continue recycling itself as long as the start lead was grounded. A minimum of two relays is needed in this example. These have four possible combinations, and every one must be used since different combinations must be established in each of the intervals 2, 3, 4, and 5. A sequence with two secondary relays is shown in Fig. 8-21 with a primary relay controlled by the start lead. Some practical difficulty might be experienced, however, in selecting relays which are slow enough in operating to give sufficient duration to the intervals. A sequence such as that of Fig. 8-22 might prove more 2 PRIMARY RELAY

*

SECONDARY RELAYS

OUTPUT TONE

*

3

4

5

ST

I

..

- -

BOTH RELAYS X ANO Y MUST BE SLOW ACTING TYPE

Fig. 8-22 Another Sequence of Secondary Relays for Fig. 8-15

Design of Sequential

Relay Circuits

165

desirable, where interval 2 is formed by three relay operating times added together, and intervals 3 and 4 are both generated by relays releasing. The choice of an operating sequence of secondary relays, when several sequences are satisfactory, may affect the complexity of required control paths. The complexity of these paths can usually be judged by inspection, then making a tentative design of the circuits for several optional sequences and observing how the modification of a given sequence will alter the corresponding circuit. Simple paths for output circuits will usually result if the operated intervals of secondary relays correspond to closures of output circuits. It may be desirable in some cases to arrange the sequence to give simple output paths at the expense of added complexity of secondary relay control circuits. In developing an operating sequence, it is often possible to use a secondary relay to perform several functions by reoperating it at several points in the sequence. This practice will reduce the number of relays required, but it has a tendency to complicate the circuit since all conditions for operating and holding the relay must be represented in the control paths. An overall improvement is often obtained in these cases by using a larger number of secondary relays controlled by simple contact networks. In the design of combinational circuits, it is frequently found possible to reduce the required contacts of certain input relays to a single make-contact and thereby eliminate corresponding relays. The same situation exists with the primary relays in sequential circuits. If a sequential circuit has been designed on the basis of a primary relay per input lead, the attempt should always be made to select secondary relay operate and release intervals which will permit elimination of one or more primary relays in this manner. A more fruitful approach in many cases, perhaps, is to start the circuit design with the idea that no primary relays are provided. The operating sequence diagram is laid out on the basis of single lead inputs and the attempt is made from the start to plan secondary relays that will function directly from input control or from the actions of other secondary relays. Primary relays are added only when necessary. Either of these methods should yield the same or equally satisfactory circuits. The second method merely approaches the desired result of minimum apparatus more directly. 8.6

ILLUSTRATIVE

EXAMPLES OF SEQUENTIAL CIRCUIT DESIGN

The design of a sequential circuit to meet stated requirements Will be illustrated first by the intercommunication signaling problem Which was stated in Section 8.4. This problem involves answer and

The Design of Switching

166

Circuits

disconnect signals to be given at positions "A" and "B" served by the intercommunication system. The circuit will require two primary relays (A) and (B) which operate when keys are operated at the corresponding positions. There are four alternative sequences of events which were illustrated in the diagram of Fig. 8-14. Two conditions occur where the same input combinations must produce different output conditions depending on the sequence of input events: the condition of (A) operated and (B) released must sometimes light the ANS-B lamp and at other times light the DIS-A lamp; likewise, (A) released and (B) operated sometimes lights ANS-A and at other times lights DIS-B. From a study of the sequences it is seen that the ANS-lamp signals always occur before an interval in which both (A) and (B) are operated and DIS-lamp signals always occur after such an interval. Therefore a single secondary relay which distinguishes between ANSand DIS- conditions should be sufficient, since the combinations of input signals can indicate the position in which one or the other of the lamps is to light. A secondary relay (X) can be inserted in the sequence as shown in Fig. 8-23. Note that this relay must return to normal in each of the normal intervals in order that one of the alternative sequences will not influence the sequence which follows. From inspection, relay (X) must operate when both (A) and (B) are operated, and hold as long as either (A) or (B) is operated. The circuit is:

f(X) = (A+ B)(X + AB) From the combinations existing during intervals 2 and 6 of the sequence diagram, Fig. 8-23, the circuit for the ANS-B lamp is: f(ANS-B)

=A+

B' + X'

This circuit will light the ANS-B lamp when (A) operates and extinguish it when (B) operates. The circuit may be simplified by reconsidering the original requirements. It will be entirely satisfactory from an operating standpoint if the ANS-B lamp is not extinguished until the secondary relay (X) operates. Thus the ANS-B lamp can lightat any time when (A) is operated and (X) is not operated, or: f(ANS-B) = A + X' This circuit does not strictly follow the sequence diagram of Fig. 8-23, since the closure of the ANS-B circuit path in interval 2 will extend into interval 3 to the point where (X) operates. The simplified circuit might have been obtained by including the combination existing in the first part of interval 3 in the symbolic expression as follows: f(ANS-B)

=(A+

B' + X')(A + B + X') =A+

X'

From an examination of the sequence diagram it will be seen that the circuit path {A + X') will be closed (or the condition of {A) operated and {X) released will exist) momentarily at times other than those

Design of Sequential 2 PRIMARY RELAYS

Relay Circuits 3

167 10

4

II

12

A

B

I3*

14

15

16 17*

SECONDARY X RELAY

-

ANS-A '---

ANS-B

----

'---

OUTPUTS i---

DIS-A

-

DIS-B

i---

* NORMAL Fig. 8-23

'---

INTERVALS

Sequence for Signaling Circuit of Fig. 8-14 with Secondary Relay Shown

specified for lighting the lamp, for example in 11 after (A) operates and before (X) operates. relay, its operating time will not be sufficient light to the point of visibility before the circuit ing this as satisfactory, similar circuit paths remaining lamps: f(ANS-A)

:;: B + X'

f(DIS-A)

:;: B' + X

f(DIS-B)

= A'

the first part of However, if (X) to permit the path is opened. can be developed

interval is a fast lamp to Acceptfor the

+X

The various paths may be combined in several is shown in the schematic diagram of Fig. 8-24.

ways, one of which

A second example will illustrate some of the hazards to the proper action of secondary relays which are sometimes encountered in circuit design. During the development of a complex relay circuit, it may not be worthwhile to consider completely the effects of every temporary path closure that may possibly occur while relays are in the process of operating or releasing. However, an analysis of these temporary closures must be made at the conclusion of design to determine whether or not they will be hazardous to proper circuit operation. A pulse-frequency divider circuit is required which will receive a series of ground pulses on a single lead and deliver output pulses at half the rate of input pulses. This circuit has many practical applications. For example, it can indicate whether the number of pulses in a series is odd or even, and it may be used as a binary counter. In solving this problem it is necessary first, to determine the number of

168

The Design of Switching

Circuits

secondary relays required; next, select a suitable operating sequence; and then, design the circuits for controlling these relays. Additional circuit paths to perform required output functions can finally be developed by combinational methods. The requirements of this problem are indicated in the sequence diagram of Fig. 8-25, in which a lead O is shown grounded during and after the odd pulses and a lead E during and after the even pulses. It will be satisfactory for the normal condition before the first pulse to give an even indication. At least two secondary relays are required. Although one relay operating through intervals 2 and 3 will provide the proper output, no control combinations for such a relay are available, since the relay must be operated and released by the same input combination. Designating the two secondary relays (W) and (Z), the sequence POSITION A

POSITION

SPEAKER ANO TRANSMITTER

8

SPEAKER ANO TRANSMITTER

KEY A

KEY

8

!_

.l

DIS-A

OIS-8

ANS-A

ANS-8

I

p ___

_rH

.__~----~

Fig. 8-24 Intercommunication

Signaling Circuit

Design of Sequential

169

Relay Circuits 2

3

4

5

INPUT

E OUTPUT

0

Fig. 8-25

Requirements

of a Pulse-Frequency

3

2 PRIMARY RELAY

SECONDARY RELAYS

Fig. 8-26

Operating

4

Divider 5

p

w

z

Sequence for Secondary Relays of a Pulse-Frequency

Divider

of Fig. 8-26 is the only suitable sequence with the trivial exceptions of changing the starting point of the cycle or interchanging relay designations. From the sequence diagram, control paths for (W) and (Z) can be developed as follows: Operate

Path for (W) = P + Z'

Hold Path for (W) = P' f(W) = (P + Z')(W + P') =PW+ Operate

P'Z'

Path for (Z) = P' + W

Hold Path for (Z) = P f(Z)

= (P' + W)(Z + P) =PW+

P'Z

The term PW may be made common to both circuits as shown in Fig. 8-27 A. This can be further simplified by modifying f(W) and f(Z)to f(W) = PW+ (P' + Z)Z' f(Z) and making P' a bridging

= PW+ (P' + Z')Z

path between Z and Was shown in Fig. 8-27B.

The Design of Switching

170

Circuits

PW-EONTi-U-IT_Y_!~,--z --

TRAN[FER

l

z•·------W B

A

Fig. 8-27 Control Paths for a Pulse-Frequency

Divider

In this form, when relay (Z) operates, the path of relay (W) is shifted from (PW+ Z') to (PW+ P' + Z), so that a continuity transfer is required on Z. Similarly, the shift from the operate to the hold path of relay (Z) on its operation requires the continuity transfer. This configuration is shown in Fig. 8-28. The hazards of this circuit are due primarily to contact stagger on the (P) relay resulting in time differences in the action of the make and the bi:eak contacts of this relay. If the contact adjustment of relay (P) is such that the make P closes before the break P' opens, a hazard will occur when the first pulse operates relay (P) in interval 2. A path will be closed momentarily through P + Z' + P' which will cause (Z) to start to operate falsely. If the stagger time of the P and P' contacts is sufficient, relay (Z) may succeed in closing its locking contact and operate fully. A second hazard occurs at the beginning of interval 4 when (P) operates on the second pulse. If P' opens before P closes (opposite of first hazard) and if (W) is very fast in releasing, the holding path for relay (Z) through contact W may open before the path through P closes, thus permitting (Z) to release sufficiently to open its locking contact and thereby fully release. This hazard seems safe unless the contact stagger on (P) is excessive. A third hazard condition occurs on the release of the (P) relay following the second pulse when the circuit restores to normal. At this time, if P' closes before P opens. (same as first hazard condition), a momentary operating path through P + Z + P' is closed to the winding of relay (W), which may operate falsely and lock.

CZ)

Fig. 8-28 Schematic Diagram of Fig. 8-26

Design of Sequential

Relay Circuits

171

Since either sequence of the operation of the P and P' contacts produces a hazard, any solution will be a compromise. The best arrangement in this case is to insure by adjustment or construction of the contact assembly of relay (P) that the break contact P' opens before the make contact closes when the relay operates. This will eliminate the first and last hazards. The second hazard can be overcome by insuring that the release time of relay (W) is sufficient to cover the time between the opening of P' and the closure of P which occurs when relay (P) releases.

p

_L

D

(Z)

(W)

I Fig. 8-29

Two-Relay

Pulse- Frequency

Divider

A circuit without operating hazards, which is often used as a pulse-frequency divider, is shown in Fig. 8-29. This circuit uses shunt control paths and operates directly from ground pulses without a pulse repeating relay. The acting sequence is the same as in Fig. 8-26, and the operation is as follows: When the first pulse is applied, relay (W) operates through its own break contact and immediately transfers through the make-before-break spring arrangement to a direct locking ground. During the first pulse, the relay (Z) is prevented from operating by the pulse ground appiied through its break contact Z' as a shunt. When the first pulse terminates, re!ay (Z) operates in parallel with (W) which continues to hold to direct ground. When the second pulse starts, ground from the pulse lead through a make on (Z) shunts (W), causing it to release. Relay (W), in releasing, transfers the holding path for (Z) back to the pulse lead. When the second pulse ends, relay (Z) releases, returning the circuit to normal.

In this circuit the hazards are overcome by using, in effect, a single make contact corresponding to the input pulse, and hence difficulties due to contact stagger on the pulse-repeating relay (P) are avoided. Since the opening of the pulse ground must cause (Z) to operate and the closure of this ground cause (W) to release, the use of shunt control paths is indicated. The control paths for (W) and (Z) in Fig.

The Design of Switching Circuits

172 2

3

4

5

INPUT P

fs (W)

SECONDARY RELAYS

w

Is (Z}

z Fig. 8-30 Sequence of Direct and Shunt Control Path Closure

8-29 may be expressed as follows, where fd represents paths and f 5 represents shunt control paths:

direct

control

fd(W) = (P + W') W

fs(W)

=P

+

Z

fd(Z) = (P + W') W f 5 (Z) = P + Z'

The intervals during which these paths are closed are shown in the sequence diagram of Fig. 8-30. The circuit might be developed by recognizing the need for shunt control, and developing the sequence diagram for direct and shunt control paths. The paths can then be developed individually and made properly disjunctive so that only a single contact, P, is required. This example illustrates that considerable ingenuity is often needed in the solution of circuit problems, and that it may be necessary to try several approaches to a given problem before a suitable circuit is found. 8.7

APPENDIX:

DEVELOPMENT OF AN OPERATING FOR A GROUP OF RELAYS

SEQUENCE

A systematic method of developing suitable operating sequences for a group of relays can be of considerable help to the circuit designer• The criterion for a suitable sequence will be that the change in combinational state at each successive interval of the sequence will be produced by the operation or release of but one element of the group. Hereafter, such sequences will be known as legitimate sequences. If

173

Design of Sequential Relay Circuits A

Fig. 8-31 Representation States of One Relay

of

The possible states which a single relay can occupy can be represented as in Fig. 8-31. The two ends of the line represent the operated and released condition of the relay, and a change in state is specified by a shift from one endof the line to the other. The concept of sequence has little meaning in this case.

A

□

AB

B

AB

A

B

Fig. 8-32 Representation States of Two Relays

of

AB

J

ri

i

there is a choice of sequences, the optimum sequence in a particular case is determined by circuit considerations.

Fig. 8-33 Representation of States of Three Relays

The four possible states of two relays can be represented by the corners of a square as shown in Fig. 8-32. Change from one state to another can take place only along the sides of the square; and, since actual sequence is not affected by interchange of designations, there is only one possible legitimate sequence, shown on Fig. 8-34. The four states of the two relays can also be represented by the line array of points also shown on Fig. 8-32, and the sequence can be determined by motion along the line. The line can be considered as folded back on itself so that a step from the B state to the null state, or vice versa, is permissible. 2

3

4

5

6

7

A

a C

2

4

A

A

B

B

C

Fig. 8-34 LeglUmate Sequence for Two Relays

Legitimate

Fig. 8-35 Sequence for Three Relays

8

The Design of Switching

174

,,--

-

AB,,,...,-;--,,

A

I

'

i+

2

...... -, \ \I

I

.!.

.!.

~

I I

:

....

C

3

•J

I

I I

I

I

I

!ti, l•.!.. j

I

.!Q.

CD

I

\ ,_, D

~

I I

..!!..,

I

I

Circuits

I I

I

121

.!!

I

.!!

j

__,,, .,,' - ...._____

I

Fig. 8-38 Representation al states or Four Relays

The eight states of three relays are represented by the corners of a cube, as shown on Fig. 8-33. Motion is restricted to the edges of the cube, and two complete legitimate sequences, indicated on Fig. 8-35, can be developed. The simpler representation of a planar array of points is also shown on Fig. 8-33. In this array, the state of a point is determined by the co-ordinate values. For example, the lower right point represents BC. Again, the array can be considered as folded back on itself so that a step from B to null, or BC to C, or vice versa is allowable. The geometrical representation of the sixteen states of four relays is a four-dimensional cube whose portrayal will not be attempted here. However, the planar array of points corresponding to this cube is drawn as on Fig. 8-36, and from this any desired sequence can be obtained. The same rules as previously described apply to motion through the array, with the addition that the top and bottom as well as the sides can be considered as folded back for the purpose of indicating allowable steps. This permits motion as shown by the dashed lines of Fig. 8-36, for example. One of the several different legitimate sequences possible with four relays is shown on Fig. 8-37. This was obtained by motion through the array as indicated by the numbered arrows on the figure. 2

3

4

5

6

7

8

9

10

II

A

8 C

D

Fig. 8-3'1 A Sequence ror Four Relays

12

13

14

15

16

Design of Sequential

Relay Circuits

175

,,------------------

--- ....., c---...-----

c---------. co---------.

co---...-----

o--------.. Fig. 8-38

',

o-------'-------------------------

Representatlon

~/

of States of Five Relays

The discussion up till now has been based upon the development of complete cycles. However, the same arrays can obviously be used for obtaining any desired partial cycle, or for developing sequences in which the same state is repeated one or more times. The expansion from the array for four relays to that for five is illustrated in Fig. 8-38. representative of a five-dimensional cube. The two arrays of Fig. 8-38 must be visualized as stacked one above the other with point E connected to the null point by a line, point EA connected to point A by a line, etc. This is indicated by the dashed lines of the figure. Rules for developing sequences are a natural extension of the previous methods. A 64-point array showing the states of six relays consists of four 16-point arrays stacked one above the other. The first two are the same as the arrays of Fig. 8-38, and, if the second one is designated the E array, the last two are the E F array and the F array, in that order. Although large numbers of different sequences for particular purposes can be developed for four, five, six, and more relays, there is a general-purpose sequence ·which may comprise any number of relays, in which each additional relay can be entered in a fixed, repetitive pattern. This is shown for three relays at the top of.Fig. 8-35, and for four relays on Fig. 8-37. The method of extension to additional relays can easily be seen.

PROBLEMS 8-1

FOR CHAPTER

8

(a) Analyze the sequence chart at the top of page 176 to determine the minimum number of relays which must be added to distlnguish between each appearance of input K. Relay (A), (B), and (C) are externally controlled.

(b) Show a circuit (A), (B), and (C).

to control

any added secondary

relays using contacts on relays

176

The Design of Switching Circuits 2

K

.......

4

3

5

6

7

8

9

10 II

12 13 14 15

16 17

18 19 20

......

-----

- - - - - - -

RELAY(Al RELAYIB) RELAYIC)

8-2

I

Design a circuit to ground a lead a during interval 7 and a lead b during interval 2, as shown In the sequence diagram below. Relays (A) and (B) are externallycontrolled primary relays; a minimum number of single-winding secondary relays may be added If necessary. Avoid continuity-transfers. I

2

34

56

7

8

RELAY (Al RELAY (Bl

-

LEAD a LEAD b

8-3

Four relays act In a cycle of twelve (12) Intervals as shown In the adjacent table. The unused combinations are assumed invalid and will never occur. Develop circuits on the contacts of these relays as follows: (Eight springs)

(a) A circuit closed during Intervals 2, 3, 6, and 7.

(Six springs)

(b) A circuit closed during Intervals 4, 5, 6, 7, 8, 9, and 10. (c) The circuit for controlling relay (B) as an internally (A), (C), and (D) being externally controlled.

controlled

relay; relays (Nine springs)

(d) The circuit for controlling relay (C) as an internally controlled relay; with relays (A), (B), and (D) being externally controlled. (Eleven springs)

A B C D 1

8-4

Design a circuit which responds to sequential input ground conditions on leads A, B, and C In the diagram below by grounding lead X or Y as indicated by the diagram. The four cycles shown are alternative. (This can be done with either four or five relays, the four-relay solution being quite difficult.) (I)

A

8 C X

y

(2)

(3)

(4)

------1

2

0

3 4 5 6

0

0 0 1 0 0 0 0 0

0

7

0

8 9 10

0 0 0 0 1 1 0 0 1 0 0

11

0

12

0 0

Design of Sequential 8-5

Relay Circuits

177

A model railroad track layout consists of three loops A, B, and C and four switches W, X, Y, and Z arranged as shown below. The switches are controlled by electromagnetic solenoids arranged so that the track Is switched to the curved spur as long as the solenoids are energized, and restored by a spring to the straight section when the solenoids are de-energized. Three track sections a, b, and c, are each provided with a single make contact which closes while the train passes over It, and which may be used to control relays for actuating the switches. The switch solenoids have no contacts. Arrange a relay circuit so that a train starting at point p and moving in the direction of the arrow will first make a complete circuit of loop A, switch to loop C, then to loop B, back to loop C again, and then to the starting point, and thereafter repeat this action indefinitely. (This can be done with two relays.)

a b

-

A

B

X

y

p

C C

8-6

A relay circuit is controlled by two keys (A) and (B), each with a single makecontact, one spring of which is grounded. This circuit grounds a lead s under the following conditions: When either key alone Is operated, ground is placed on lead s. Upon the release of the key, the ground is removed from s. However, If after one key is operated and s is grounded, II the second key is operated before the first Is released, the release of either key will remove ground from the lead s. Design such a circuit with the additional requirement that no relays shall remain operated when both keys are in an unoperated condition. Avoid shunt-down conditions II possible. (This can be done with three relays.)

8-7

In order to protect the power tubes of a particular radio transmitter, the plate heated. voltages must not be applied to the tubes until the cathodes are sufficiently A relay circuit Is to be devised which will supply current to all filaments when the transmitter switch is turned on, will supply plate voltage to the driver stage not less than 30 seconds (and not more than 45 seconds) after the filaments are energized, and will apply plate voltage to the power ampli!ier not less than 45 seconds (and not more than 1 minute) after the filaments are energized. Times need be accurate to no more than one second. To obtain the necessary timing Intervals, a 115-volt a-c synchronous motor timer from Its back to its front contact for Is supplied which switches a transfer-contact one second every 15 seconds. This timer should be energized only during thesequence described above, and may rest on either the front or back contact when the switch is turned on. Design the required relay circuit. (This can be done with three relays.)

8-8

A circuit controlled by two single-make non-locking keys, A and B, has as outputs a red lamp and a green lamp. The circuit always assumes a normal condition with

178

The Design of Switching

Circuits

no relays operated and no lamps lighted when the two keys are released. Starting from a normal condition, the first key operated, either A or B, lights lamp G. With the first key still operated, operating the second key, either B or A, will light lamp R. Thereafter, as long as either or both keys remain operated, the lamp first associated with a particular key will follow the condition of that key: i.e., it will light when the key ls operated and go out when the key is released. (This can be done with two relays.) 8-9

A rotating shaft carries a single grounded brush which makes contact with three stationary commutator segments arranged symmetrically around the shaft. A relay circuit is required which will indicate the direction of shaft rotation by lighting one of two lamps, designated CW and CCW, for clockwise and counterclockwise respectively. The shaft may reverse its direction at any time. Assume that the shaft ls driven so that a brush contact closure is 0.25 second, and that the open time between the brush leaving one segment and reaching another is 0.25 second. When the shaft changes direction, the output indication must change as quickly as possible, at most within 2 seconds. (This can be done with four relays; to indicate change within one revolution may take six relays.)

Chapter 9 ELEMENTARY

SWITCH CONTROL cmCUITS

Some of the functions which switching devices may be called upon to perform can be efficiently implemented by the aid of rotary and relay-type switches. This is true, for example, of the task of establishing connecting linkages through telephone switching systems - a major application for the crossbar switch, since the ratio of connectable contacts to electromagnetic control parts for this device is high. Similarly, the mechanical construction of a rotary selector switch makes its use as a sequence controller highly attractive where one circuit is to be connected, in turn, to each of a number of other circuits. Switches find many other applications throughout the field of automatic control systems.

As is found in the case of relays, certain basic switch circuits repeatedly make their appearance in the switching art, the particular configuration of circuit depending upon the type of switch employed and the function it is to perform. Switches occur in a wide variety of kinds, each of which demands particular consideration as to the best method of control. In this chapter only the construction, operation, and control of the so-called rotary selectors and relay-type switches will be discussed. Although many other types of switch mechanisms are extant, the operation and control of these mechanisms should present no difficulties if the rotary selector and relay-type switches and associated control circuits are thoroughly understood.

9.1

ROTARY SELECTOR SWITCHES: MECHANICAL CONSIDERATIONS

Rotary-type selector switches consist, primarily, of arcs of terminals over which associated wipers pass. An electromagnet mounted on the switch assembly provides power to move the wipers from one terminal position to the next, each separate energ_ization and de-energization cycle of the magnet causing the wipers to move one position. There are two basic types of rotary switches: forward-acting or direct driven switches, which step from one terminal to the next terminal on the energization of the magnet; and back-acting or spring driven

- 179 -

180

The Design of Switching

Fig. 9-1 Back-Acting 22-Polnt Rotary Selector

Circuits

Switch

switches, that step on the de-energization of the magnet. The control magnet of either type of switch is known as the "step magnet". When the step magnet of a forward-acting switch is energized, a pawl coupled to the magnet armature is forced against the teeth of a ratchet wheel on the shaft supporting the wipers, causing the shaft to rotate through a small angle, thus moving the wipers from one terminal to the next. A detent engaging the ratchet wheel insures that the wipers remain on the terminal just reached when the magnet is de-energized. In the case of a back-acting switch. energization of the step magnet pulls a pawl away from the ratchet wheel on the wiper shaft, against the force of a spring attached to the frame of the switch. When the magnet is de-energized, the pawl is pulled back by the spring, engaging a tooth on the ratchet wheel and advancing the wipers a single step. Some rotary switches may be caused to step continuously in the same rotary direction over the same set of terminals, whereas others, after stepping the ii:..wipers over the associated arcs, must be returned to a normal position before the wipers can again be moved over the arc terminals. These two types of switches are designated "non-homing" and "homing", respectively. Switches of the homing type are normally equipped with a second magnet, a release magnet, which removes a restraining detent from the ratchet wheel and allows a spring to restore the wipers to the starting position. Wipers arc terminals

may be either of two types: bridging, in which adjacent are short-circuited by the wiper as it steps from one to

Elementary

Switch Control

Circuits

181

the other; and non-bridging, in which before it makes contact with the next.

the wiper

leaves

one terminal

Contact springs are often provided on the step magnet to indicate that the magnet is fully energized. Also, for those switches which must be returned to a normal position, off-normal contacts are employed to indicate when the switch is not in the starting position. These contacts are in their normal or unoperated state only when the switch wiper assembly is in a so-called "home" position. Illustrated in Fig. 9-1 is a back-acting rotary switch consisting of six arcs of 22 terminals each. The wipers are double-ended so that, when one wiper end has passed over the half-circle of 22 terminals, the other end is in position to start stepping over the same 22 terminals. Occasionally, single-ended wipers mounted in pairs staggered 180 degrees apart are utilized on a similar switch so that two adjacent arcs of 22 terminals may be employed as a continuous bank of 44 terminals. The stepping magnet, on the right-hand side of the picture, is equipped circuit with a break contact. The switch may be driven by external pulses at a rate of up to 25 or 30 steps per second, and a rate of 50 or 60 steps per second may be realized if the switch runs under selfinterrupted control as discussed later in the chapter. The resistance of the step magnet winding is normally low (200 ohms or less) in order to meet mechanical power and speed requirements. Circuit drawing conventions for such a switch are pictured in Fig. 9-2. The bridging and non-bridging wipers are indicated by BDG and NBG respectively, and the numbers in the center of the wiper circles are the arc numbers. It is seen that, in actuality, there is a twenty-third terminal on the 22-position arc which is permanently connected to.the associated wiper. This terminal merely serves as a means for joining external wiring to the wiper. 15.0 oooe 00

20.

0 0

22~~

Fig. 9-2

10

150 oo

000

o• 0

0.5

O 0

01

20.

O 0

0

10 •o

0

oo • 5 0

0

01

22:~-\ WIPER

WIPER

CONNECTION

CONNECTION

Drawing Conventions for 22-Polnt Rotary Selector Switch

A forward-acting ten-terminal two-arc switch is illustrated in Fig. 9-3. In addition to the stepping magnet, this switch is furnished With a release magnet, shown in the lower left corner, and off-~ormal springs, in the upper right corner. In the normal position, the wipers stand in the position just preceding the first terminal. The switch may be driven at speeds up to 25 steps per second. As in the previously

The Design of Switching

182

Fig. 9-3 Forward-Actlng

described switch, generally low.

Circuits

10-Polnt Rotary Selector Switch

the winding resistance

of each control

magnet

is

As a wiper passes over its arc, a second wiper connected electri.cally and mechanically to the first moves along a copper strip placed beneath the arc. This copper strip provides an electrical path to the first wiper and, for this purpose, is equipped with a wiring lug. Fig. 9-4 shows drawing conventions for this switch. In the text, both types of switches will be shown by a somewhat simpler schematic symbol, illustrated by Fig. 9-9. Unless a homing switch is specified, it should be understood that the wiper can step directly from one end terminal of the arc to the other.

Flg. 9-4 Drawlng Conventions for 10-Polnt Rotary Selector

Elementary

9.2

Switch Control

CONTROL

As step for magnet. one step

Circuits

OF ROTARY

SELECTOR

183 SWITCHES

discussed above, rotary selector switches are advanced one each cycle of energization and de-energization of a stepping The circuit which insures that the switch takes one and only per external control action is termed a "stepping" circuit.

Switching mechanisms are frequently required to step continuously, repeatedly recycling, when only a single external condition is present. This type of switch operation is known as "running" and the controlling circuit is then a running circuit. If a separate circuit action is employed to stop the running process at a specific switch position, the circuit providing the means for stopping the switch is referred to as a "halting" circuit. Finally, it is often necessary, after a certain sequence of switch operations, to return the wipers of that switch to a starting or normal position. This is accomplished by a "restore-to-normal" circuit which may either step the switch to the required position or operate a release magnet provided on the switch. Any one of these control circuits .is composed of one or more of relays or other control devices external to the the following ~lements: switch; contacts activated by the stepping magnets of the switch; and terminals of the switch arc and the associated wiper. The sections to follow indicate a number of methods by which rotary switches may be controlled in the stepping, running, halting, and restoring functions. EXTERNAL CIRCUIT

Flg. 9-5 Dlrect Stepping

9.3

SWITCH STEPPING

CIBCUITS

The most convenient method of stepping a rotary switch from one terminal to the next is that of directly energizing the step magnet until the pawl has reached its fully operated position and then de-energizing the magnet, allowing the pawl to return to rest. This may be accomplished by such a circuit as that shown in Fig. 9-5, a direct stepping process in which the step magnet is energized by each closure of the make-contact on relay (P) in the external circuit. If the switch is of the forward-acting type, it will step at the beginning of the ground pulse; whereas a back-acting switch will step after the step magnet is deenergized.

The Design of Switching

184

Circuits

EXTERNAL CIRCUIT

_[

Ont

u

(P)

Fig. 9-8 Control of Pulse Length: Pulse Too Short

It may occur that the ground pulse applied by the external circuit to the switch is not of sufficient duration to operate the step magn~t fully, and hence, might fail to step the switch. In this case it is necessary to lengthen the energizing pulse. This is effected by such an arrangement as the circuit of Fig. 9-6. The input ground pulse operates relay (A) which locks to the break-contact on the step magnet. Ground is applied to the step magnet until relay (A) has released. The adjustment of the contact on the step magnet influences the duration of the magnet energization in that the contact does not open until the pawl has been moved the required distance.

EXTERNAL CIRCUIT

Fig. 9-7 Control of Pulse Length:

Pulse Too Long

Conversely, the actuating ground pulse may be applied· by the external circuit to the step magnet for a considerable time. As mentioned above, switch magnets are in general of low resistance to facilitate rapid stepping, and therefore should not be energized for a time greater than is necessary to permit the switch to step. One method by which ground may be removed after the switch has stepped is illustrated in Fig. 9-7. Here, ground is applied to the switch only during the operate time of the slow-operating (A) relay. Note that the switch steps either when relay (P) operates or when relay (A) operates, depending upon whether the switch is forward- or back-acting. Relay (A) may be replaced by any timing device which performs an identical function.

Elementary

Switch Control

Circuits

185

EXTERNAL CIRCUIT

l__ -=-

0

+

0

~

(Pl

(Al

Fig. 9-8 Stepping Forward-Acting

Switch at Termination

of Pulse

A forward-acting switch may be made to advance in a manner similar to a back-acting switch, the wipers remaining fixed in position until the termination of the input pulse. A circuit making this possible is shown in Fig. 9-8. The pulsing relay (P) of the external circuit operates slow-release relay (A). When (P) releases, (A) remains operated for an appreciable time because of its release characteristics, closing the path to the step magnet. The forward-acting switch steps before (A) removes ground from the magnet.

EXTERNAL CIRCUIT

0

0

0

_[

nt lJ

(Pl

STEPPING

THROUGH ARC -

FORWARD-ACTING

SWITCH

A

EXTERNAL CIRCUIT

..L -

0

~t

(Pl

0

0 0

0

lib_

0 0

-

0

STEPPING THROUGH ARC -

BACK-ACTING

SWITCH

B Fig. 9-9 Stepping from a Particular

Arc Terminal

The Design of Switching

186

Circuits

In a given application, it may be necessary to step the wipers of a rotary switch from a particular position to the next. This is accomplished by applying ground to an arc terminal in that position, and connecting the step magnet to the wiper of the associated arc. The actual circuit used depends upon the type of switch employed; Figs. 9-9A and 9-9B illustrate arrangements for forwardand back-acting switches. A break-contact is utilized on the step magnet for the backacting switch to allow the switch to step while relay (P) is still operated. If the switch is in proper adjustment, this break does not open until the It is magnet is sufficiently energized to operate the pawl satisfactorily. necessary in the forward-acting arrangement that the detent pass over the ratchet tooth before the wiper moves completely off the arc terminal which establishes the stepping magnet operate path. Switches are designed to meet these requirements.

EXTERNAL CIRCUIT

_[

0

~

(RI

Fig. 9-10 Direct RuMlng, Self-Interrupted

9 .4

Circuit

SWITCH RUNNING cmCUITS

In running, it is necessary for the switch step magnet to be alternately energized and de-energized for as long as an external ground signal is applied to the input of the circuit or until the switch is halted. The most frequently used running arrangement is the "self- interrupted" type of circuit in which contacts on the stepping magnet structure are employed in the operating path. These contacts, which close or open on each activation of the magnet, are known as "interrupter" contacts since, in a self-interrupter circuit, they are used to interrupt the flow of current to the magnet. It is normally assumed that these contacts will act only when the pawl and detent are in the proper mechanical relation to the ratchet wheel to insure complete operation of the stepping mechanism. However, such factors as inertia of the armature assembly may permit the contact to be effective in a self-interrupted arrangement even though the mechanical relationship is not perfect. The basic self-interrupted running circuit, 9-10, uses a break contact in series with the magnet magnet has been energized to an extent sufficient to the energizing path is opened and the pawl assembly

illustrated in Fig. winding. When the actuate the switch, returns to normal,

Elementary

Switch Control Circuits

187

completing the cycle by reclosing the interrupter contact. This arrangement is applicable to both back- and forward-acting switches. In such a direct running circuit, some switches may attain a speed of 50 or 60 steps per second since full energization and de-energization of the magnet for each step is not necessary. Such rapid running may be undesirable, particularly if the switch is to perform functions at certain positions during the running process. Han additional relay is included in the circuit, the acting time of this relay can be added to that of the step magnet for each stepping cycle. To illustrate, two circuits are shown in Fig. 9-11, one employing a make interrupter contact and the second, a break. In Fig. 9-llA, the step magnet operating path is not opened immediately after the magnet has been energized, but instead, remains closed during the operating time of relay (A). After the operating path is opened, it is not again closed until (A) has released. H the spring load of relay (A) is light, its release time will ordinarily be greater than its operate time; and as a result the switch, either back- or forward-acting, is de-energized for a greater duration than it is energized.

EXTERNAL CIRCUIT

~ 7

,.,

~

_r,11sET t p

fl

(A)

1!1Q_

MAKE-CONTACT ON STEP MAGNET A

EXTERNAL CIRCUIT

(Bl

BREAK-CONTACT ON STEP MAGNET

B Flg. 9-11 Two Examples of Slow-Running Clrcults

The Design of Switching Circuits

188

H only a break interrupter contact is available, an arrangement similar to that of Fig. 9-11B is employed. Here, the period of magnet energization is longer than the period of de-energization. This may be a disadvantage in the case of a back-acting switch since, with such a type of switch, arc-terminal conditions are tested during the de-energization period. H, for example, a condition on a particular terminal is to inform the back-acting switch to cease running when that terminal is reached, information that the operating path is to be opened must be transmitted, and the opening of the path must occur, all during the deenergization period which follows after the switch has reached that particular terminal. To increase this de-energization period, the relay (B) in Fig. 9-11B can be chosen to be somewhat slow in operating; the circuit in Fig. 9- llA can also be made slower running by selecting a slow-release relay for (A). It is normally of advantage, however, to permit the switch to run at as great a speed as is consistent with the requirements of the circuit. An interrupter contact is not always available on the stepping magnet for running purposes, precluding the use of a self-interrupting circuit. In such instances, an interrupter circuit or pulse generator must be employed, as in Fig. 9-12A. The pulse generator may be a EXTERNAL CIRCUIT

PULSE GENERATOR

A EXTERNAL CIRCUIT

_[

0

~(Al~

B Flg. 9-12 Running with Pulse Generator

Control

Elementary

Switch Control

Circuits

189

motor-driven interrupter, a gas-tube pulsing circuit, or any other convenient pulsing arrangement. A switch controlled by a relay pulse generator is illustrated in Fig. 9-12B. The slow-acting characteristics of the relays (A) and (B) limit the speed at which the switch runs. As in the circuits of Figs. 9- llA and 9-11B, the release of relay (R) will immediately remove ground from the step magnet, preventing further running. Such direct control is desirable in that the progression of the wipers can be halted in the shortest possible time.

EXTERNAL CIRCUIT

0

BOG

Fig. 9-13

9.5

SWITCH HALTING

Direct

Halting by Absence of Ground

CIRCUITS

The presence of a running circuit in the control of a rotary switch ordinarily implies that at some point the switch must cease to run. If there is included a method of halting the switch other than merely removing the input condition causing running, this arrangement assumes the form of a halting circuit. In halting circuits, means are provided for stopping the progression of the switch in some position as indicated by the presence or absence of a potential (battery or ground) condition on an arc terminal. In all the examples to follow. the switches run to a terminal marked by the presence or absence of direct ground, it being understood that battery or some marginal condition may be substituted for the ground condition by appropriate circuit design. One of the most useful of halting arrangements consists of a selfinterrupted running circuit with the running ground passing through an arc and wiper of the switch under control. In the position or positions at which the switch is to be halted, the arc terminal is not connected to is illustrated in Fig. the external cii:cuit ground. This basic circuit 9-13. The arc terminals are protected by the use of a bridging-type wiper. In this figure, as well as in some of the others of this chapter, the switch arc is shown connected to ground and the wiper to the stepmagnet circuit. However, if it seems desirable, these connections may be reversed. The current through the switch arc can be reduced by the

The Design of Switching

190

Circuits

~ -=-IRlo......,-----,

A

EXTERNAL CIRCUIT

t•I -=-

l~L.~------,-------

B

Fig. 9-14 Relay-Controlled

Running and Halting Circuits: Halting by Absence of Ground

addition of a repeating relay placed between the wiper and the interrupter contact, with ground being applied to the step-magnet circuit through a make-contact on this added relay. Used with a bridging wiper, this relay remains operated until the desired position is reached. The relay must be fast in releasing when the ungrounded terminal is • found to prevent the step magnet from remaining energized for an additional step. • With the circuitofFig.9-13it is necessaryto employ an additional switch arc and wiper to detect that the switch has been halted. This . may be undesirable if all arcs are otherwise used to capacity; in such a case it would be convenient to provide means in the halting circuit itself to indicate switch stoppage. A halt-indicating circuit for use with a forward-acting switch is shown in Fig. 9-14A. Relay (S) is marginal in that it will release when connected in series with relay (H). The (S) relay is a part of a running circuit similar to that of Fig. 9-11B. Relay (H) is locked down until the control wiper reaches the ungrounded terminal, at which time it operates, removing the running ground from (S) and placing ground on lead wiper is a as an indication that the switch is halted. If a non-bridging employed, the dashed wiring should be incorporated to insure that (H)

Elementary

Switch Control

Circuits

191

will remain unoperated as the wiper passes between grounded terminals. A back-acting switch with bridging wiper may be substituted in Fig. 9-14A. The dashed wiring is of no assistance in this case. In all the running circuits mentioned so far, the operation of a relay (R), external to the running circuit, has provided the running ground. In the circuit of Fig. 9-14B, relay (R) must be operated and released before running can commence. Upon operation of (R), relay (A) operates and locks to the break interrupter contact of the switch. When (R) is released, ground is applied to the step magnet which runs in self-interrupted fashion until relay (A) releases. Relay (A) can release only after the step magnet has been energized, the interrupter contact has opened, and the wiper has reached an ungrounded terminal. Thus the forward-acting switch will halt on the ungrounded terminal, whereas the back-acting switch will step to the next adjacent terminal. In the circuit, the wiper need not be of the bridging type since the locking path of relay (A) through the interrupter contact covers the interval during the wiper step movement. Relay (A), however, must be fast in releasing when the ungrounded terminal is reached since the interrupter contact in a rapidly self-interrupted circuit does not long remain open. a 0

0

_r--x ....

\

/ 0

tA1~1•~

0

.Lnto•

L]

0

..-(RI

EXTERNAL CIRCUIT

~I•~

Flg. 9-15 Basic Halting by Presence of Ground

It is often more appropriate to halt a running switch on a grounded terminal since the running circuit can then be opened by the operation of a control relay rather than by its release. This is desirable since the control relay with its small contact-spring load is usually more rapid in operation than in release. Fig. 9-15 shows a self-interrupted running circuit controlled by fast-operating relay (A). This relay acts when the marked terminal is reached, halting the switch and placing an indicating ground on lead a. This is a widely-used circuit, both as shown and also in a form where the two connections to the arc and wiper of the switch in Fig. 9-15 are-reversed.

192

The Design of Switching Circuits 0 0

_rx-=-

EXTERNAL CIRCUIT

0

0 0

Fig. 9-18 Halting by Presence of Ground: Shunt Operation

Fig. 9-17 Basic Restoring Circuit:

EXTERNAL CIRCUIT,---+- ______

_[

ao'

-=-(RN)

rb

Homing-Type Switch

__,

"' Fig. 9-18 Restoring Circuit:

Fig. 9-19 Arc-Controlled

Homing-Type Switch

Homing Switch

0.N. ,,------o-.~f\

_l_

a

0

-=-IRNI

Fig. 9-20 Off-Normal Contact Controlled Restoring Circuit:

Non-Homing Type switch

Elementary

Switch Control

Circuits

193

The control relay is not vital to a ground-halted circuit, as is illustrated by Fig. 9-16. Here, the step magnet is shunted down when a grounded terminal is reached, preventing further stepping. This circuit gives no output indication of switch stoppage; moreover, considerable current is drawn through the arc and the protective resistance.

9.6

SWITCH RESTORING

CIRCUITS

When it is desired to restore a homing switch, a path, usually through an off-normal contact on the switch, is closed to its release magnet. The release magnet is thus energized and held energized until A circuit effecting this is the switch has _been restored to normal. shown in Fig. 9-17. Since the restoring ground is supplied through the off-normal contact, the release magnet draws current only until the switch is restored. circuit is of insufficient If the input ground from the external duration to insure that the switch is fully restored, a circuit similar to that of Fig. 9-18 must be employed. Otherwise the released detent may damage the ratchet wheel during the restoring movement of the switch mechanism. Care must also be taken that the step magnet is not energized during switch restoration. In Fig. 9-19, the switch is returned to normal automatically when the wiper has reached its final position on the arc. Means can be provided by the circuit shown in dotted lines for restoring the switch to its home position at any time desired. Rotary switches which do not have a mechanical frequently require, for circuit reasons, the designation terminal as "home". To reach this terminal, the switch its arc and halted at the home position by some method already discussed. The circuit of Fig. 9-13 is usually result of its speed of action.

normal position of a particular is run through similar to those preferable as a

If a non-homing switch is equipped with off-normal springs, the off-normal contact may be used in restoring the circuit to normal. As in Fig. 9-20, the switch runs in a self-interrupted manner until the offnormal contact opens. Occasionally, a set of "off limits" contacts are provided to indicate that the switch has reached the limiting position on its arc and must be returned to normal before the circuit can again function. When these contacts are actuated, the switch may be restored in a circuit similar to that of Fig. 9-19, in which the wiper and grounded arc terminal are replaced by an off-limits make-contact.

The Design of Switching

194

Circuits

Fig. 9-21 Crossbar Switch, Front View

9.7

CROSSBAR SWITCHES: MECHANICAL CONSIDERATIONS

The crossbar switch, shown in front view in Fig. 9-21, comprises a welded rectangular frame on which are mounted a number of vertical relay-like units and five horizontal bars turned by so-called selecting magnets. Each vertical unit has a "holding" magnet and ten sets of contacts in a vertical row. The holding magnet, or, briefly, hold magnet, actuates a vertical armature which is adjacent to the ten vertical sets of contacts. Each contact set, known as a crosspoint, consists of from containing three to six make-contacts, all sets of a given unit usually the same number. The internal structure of a vertical unit connects the fixed spring of each make-contact to the corresponding springs of the other sets on the same vertical unit, thus producing a permanent vertical multiple. The movable springs are brought out individually, but may be wired externally to form a horizontal multiple. Thus, a fully multipled switch would comprise a group of vertical units each connectable to any one of ten horizontal levels. A co-ordinate representation of this is shown on Fig. 9-22. 9 8 7 6 5 4 3 2

-;--;--t-t--+--+--+-1--4---4---+--+-~-~-4----1---l--L---L--L---t---t---i-+--+--+~--1--4----1----t---t---i-+--+--+~--1--4----1--

9---l---------+-

--+---+----i-+--+---1-----1-"'--~---L--

-;--;--t-t--+--+--+-1--4---4----t--t--t-t--+--+--+-1--4---4--

--t--+---i-+-+--+~-+---+----+--

1 ---t--t--t-t--+--+--+-1--4---4--

0 ---t--t--t-t--+--+--+-1--4---4--

0-4--------+0

0123456789

Fie. 9-22 Crossbar Switch, Co-Ordinate Representation

9

Elementary Switch Control Circuits

195

The horizontal select bars are rotatable in either o! two directions by two select magnets associated with each. At each vertical unit location there is, on the horizontal bar, a flexible selecting finger which, when the bar is rotated by the energization or one or its two select magnets, interposes itself between the armature of the vertical unit and the associated set o! contacts. This is shown on Fig. 9-23. I! a hold magnet operates after the select finger is in this position, the set of contacts at the crosspoint so selected ls closed. Thus, these crosspoint contacts can be operated independently of each other by a coordinate action of the horizontal bars and vertical armatures. The select magnet, when de-energized, restores the horizontal bar to its neutral position, but the operated crosspoint holds its associated selecting finger. If a select magnet is energized after a hold magnet is operated, the contacts at the selected crosspoint will not close since prior operation or a hold magnet prevents the inter-position or a select finger into any crosspoint in the associated vertical unit. There are, in general, two sizes of crossbar switches; one with 10 hold magnets and 100 crosspoints, the second with 20 hold·magnets and 200 crosspoints. Other variations with respect to the number o! holding magnets are also in use.

Flg. 9-23 Crossbar

Switch: Crosspoint Detail

On inspection of Fig. 9-23, the details o! a crosspoint become more evident. The movable contact springs are bifurcated and the contacts themselves are of precious metal, thus assuring good contact characteristics through the switch. The number of make-contacts at each crosspoint is frequently used in describing a switch. For example, a switch with three make-contacts per crosspoint is referred to as a three-wire switch .

•

196

The Design of Switching

Fig. 9-24 Wiring Terminals

for Movable Springs of a Crossbar

Circuits

Switch

The various select and hold magnets themselves may be equipped with off-normal spring combinations. These are operated directly by the associated magnet irrespective of which crosspoint is closed. Wiring lugs for the movable springs of the crosspoints are arranged for individual wiring, as illustrated in Fig. 9-24 which shows the wiring terminals for two crosspoints. The terminals are provided with staggered notched projections to facilitate horizontal strapping of corresponding terminals of horizontal groups of crosspoints. Fig. 9 -25 shows the circuit diagram conventions used for the crossbar switch. 9.8

CONTROL

OF THE CROSSBAR SWITCH

The control of the crossbar switch is very similar to that of ordinary relays except that mechanical considerations dictate certain rules which must be followed in controlling the switch. Because of the construction of the switch, once a crosspoint has been closed and is held by its associated hold magnet, no other crosspoints in the same vertical unit can be actuated. More than one crosspoint in the same vertical unit may be closed only if the corresponding select magnets are operated before energizing the hold magnet. In this connection, it must be remembered that there are two select magnets for each horizontal bar and .that only one of these should be energized at a time. The operation or release of the horizontal bars results in transient vibration of the select fingers. For this reason, time must be allowed in the circuit operation for the fingers to become reasonably stationary before attempting to operate a hold magnet, in order to insure against either missing a finger or "snagging" an extra finger when

•

Elementary

Switch Control

197

Circuits

the hold magnet operates. Since each finger of a select bar, held by a vertical armature, exerts further tension on the bar, a limit of nine select fingers held on any horizontal level may be prescribed. The hold magnets of a crossbar switch are usually energized for a considerably greater time than the select magnets, since they hold the established connection. For this reason, their windings are ordinarily of high resistance to conserve power. The operating time for a crosspoint, measured from the instant the circuit to the hold magnet is closed ·until the last contact of the crosspoint operates, varies from about 0.025 second to 0.060 second, depending upon the number of contacts at the crosspoint and upon the off -normal spring combination. Select magnets, to be rapid in operation, are ordinarily with 600-ohm coils, though smaller values may be used. generally energized only for short periods of time. Time closure of the select magnet circuit until the hold magnet may energized ranges from about 0.010 to 0:040 second.

provided They are from the be safely

To provide the necessary delay between the operation of a select magnet and that of the hold magnet, any convenient timing arrangement may be used. Select magnet off -normal contacts may also be employed in the control of the hold magnet. Very often, in large circuits using crossbar switches, it is possible to utilize the inherent time delays resulting from the performance of other functions to guarantee the necessary time interval, rather than providing specific relays for the purpose. One of the principal advantages of the crossbar switch is that it is inherently more rapid in operation than the rotary-type switch. To disconnect an input from one output terminal and reconnect it to another,

ASSOCIATE WITH CORRESPONDING VERTICAL

Flg. 9-25

Crossbar

Switch:

Circuit

Diagram Conventions

198

The Design of Switching

Circuits

the rotary-type switch must move in a progressive manner over all intervening terminals. In accomplishing an identical task. the relay circuit associated with the crossbar switch can control the disconnection and re-connection, one immediately following the other. Another factor contributing to the speed of crossbar switch circuits is that no connection through the switch exists until a hold magnet is operated. This allows the select magnets to be operated by the controlling circuits at some time before the connection is to be made. The select magnets of the switch may follow the actions of control circuits with no effect on their associated crosspoints until such time as a hold magnet is energized. Although, generally, only one select magnet on a switch is operated to prepare for a connection, in certain applications two crosspoints must be closed to provide a path through the switch. thus usually necessitating the energization of two select magnets. An example of such a case is discussed in the following paragraph. OUTPUTS

7

0

15

-

.._______

8

--oo INPUT

l---=J~o9

~

INPUT

~ Flg. 9-26 Spllttlng Crossbar Swltch to Increase

the Number of Outpits

Elementary

Switch Control

Circuits

199

In normal operation the vertical unit of a crossbar switch is equipped with a single input with access to any one of ten outputs. By holding the number of leads in each output to three or less, the number of outputs may be increased. This can be done, for example, with a sixwire switch. For this purpose the vertical unit of six vertical multiple ~traps is divided into two groups of three straps each, as illustrated in Fig. 9-26, thus effectively providing the unit with twenty crosspoints, each consisting of three make-contacts. The vertical unit is equipped with one three-wire input which is connected to the "0" horizontal level that is associated with the first group of vertical multiples, and is in addition connected to the "1" level associated with the second group of vertical multiples. The crosspoints of the "l" horizontal level of the first group and the "0" level of the second group are left vacant. The remaining sixteen crosspoints of the unit each provide an output connection as shown in the figure. The outputs can be multipled across all vertical units. By this device, each unit has been altered to give sixteen outputs rather than the standard ten at the expense of reducing to three the number of leads to be connected. For any connection, two select magnets must be operated as well as a hold magnet. To illustrate, if input leads "0" are to be connected to output leads "12", select magnets (1) and (6), and hold magnet (0), are energized. The particular hold magnet to be operated depends on the input leads, and the select magthat the nets depend on the output desired. It should be remembered crossbar switch has no directional sense, so that the terms input and output can be interchanged in the above example. Since the design of a crossbar switch control circuit depends primarily upon the function which the switch is to perform and its relation to other associated circuits, detailed discussion of specific control arrangements is- not included at this juncture. However, no great difficulty in the understanding of such circuits should be experienced because of the previously mentioned similarity of the crossbar switch to a relay arrangement.

PROBLEMS

FOR CHAPTER

9

9-1

A clrcult ls req·Jlred to count up to 100 pulses from a key K with a single makecontact, giving a continuous output _indication by grounding one of 100 leads. The circuit Is to employ a 22-point, back-acting, non-homing rotary switch and need not be recycling. The circuit should be active only when an ON-OFF switch Is turned ON, and should restore to normal at any time when the switch Is turned OFF. (This can be done with four relays in addition to the switch.)

9 -2

A 10-point forward-acting homing-type rotary switch Is to be used ln an application ln which it Is necessary that the wipers normally rest on the No. 5 terminal. When a non-locklng key K ls operated momentarily, the wipers run through terminals 6,

The Design of Switching

200

Circuits

7, 8, 9, 10, 1, 2, 3, and 4, ln order, and halt at terminal 5. On alternate rounds, the output leads connected to terminals 6, 7, 8, 9, and 10 are to be grounded in reverse order: that is, for odd rounds the order or grounding output leads is 6, 7, 8, 9, 10, 1, 2, 3, 4 and for even rounds, the order is 10, 9, 8, 7, 6, 1, 2, 3, 4. The switch is furnished with a step magnet and a release magnet, neither equipped with contacts. However, switch off-normal springs may be specified. Two terminal arcs are available on the switch. Design a suitable switch.

clrcult.

(This can be done with Clve relays

in addition

to the

9-3

(a) An 11-polnt three-arc back-acting rotary switch Is used as a translating device. A ten-button key set grounds two leads out of five designated 0, 1, 2, 4, 7, in an additive two-out-of-five code as shown below. A sixth lead from the keyset ls grounded to indicate that a key is depressed. One of the switch positions is designated as the home or normal position. There is no release magnet; the switch ls of the nonhoming type. Design the switch circuit to light one or ten (10) lamps indicative of the key which is depressed. Arrange the circuit to return the swltch to the normal position after Leads Key any depressed key Is released. Allow only the Grounded Operated desired lamp to light during the circuit operation. 0, 1 1 (This can be done with three relays.) 0,2 2 (b) Using a six-arc swlfch ln place of the three1, 2 3 arc switch, arrange the circuit or part (a) to , 0, 4 4 check for exactly two-out-of-five leads grounded 1, 4 5 without employing a 2/5 or 3/5 symmetric con2,4 6 tact configuration on any or the relays. An In0,7 7 correct indication should cause an alarm signal 1, 7 8 to lock In and prevent any numerical lamp In2,7 9 dication. (This can be done with one additional 4,7 10 relay.)

9-4

When an ON-OFF switch ls turned ON, a back-acting 22-point non-homing rotary switch ls to follow the operation or a key K, equipped with a single make-contact. The switch is to step on each depression or the key and again on each release of the key. Uthe key ls held operated for an Interval longer than approximately 1/4 second, the switch is to restore to its normal position and to ignore the subsequent release of the key. After the switch has restored to normal, and after this final release or the key, the switch must again be free to act as specified above. Design the required circuit.

9-5

(This can be done with two relays.)

A radio receiving system Is to be installed. There are to be ten loudspeakers distributed throughout the premises, but only one of these Is to be connected to the central receiving unlt at any one time. A 22-point rotary switch ls to be used as the connecting device, cutting through the two output leads from the receiver to the desired speaker. Characteristics

of Apparatus:

Speaker Impedance: 500 ohms Receiver Output Impedance: 500 ohms Rotary Switch: 22-point, back-acting, non-homing. Three terminal arcs. One break-contact on step magnet. per key. Selector Keys: Non-locking, two make-contacts

Elementary

Switch Control Circuits

201

The particular speaker to be driven is selected by operating one or ten non-locking keys. The connection thus established to this speaker Is maintained until some other key is operated. During the acting time of the switch no momentary connection should be made to any speaker. However, the load impedance presented to the receiver during this acting time should be the same as that provided when a coMecThe switch and its control relays should draw no current except tion is established. when establishing a new connection. (This problem 9-6

can be done with two relays.)

Design a control arrangement for a select and hold magnet of a crossbar meet the following requirements:

switch to

(a) Action is initiated by closure of a single make contact which remains only long enough for the select magnet to operate fully.

closed

(b) The circuit shall make sure that the select magnet has operated before starting timing for hold magnet operation. (c) The operate operation.

time of three relays shall intervene between select and hold magnet

(d) The circuit shall make sure that the hold magnet has locked operated to a separate ground before releasing the select magnet and restoring to normal. (e) The hold and select magnets are each furnished contact. (This problem

can be done with three relays.)

with a single oil-normal

make-

Chapter 10 ELEMENTARY ELECTRONIC SWITCHING CIRCUITS

Electronic tubes and the semi-conductor elements, varistors and transistors, can be used as two-valued devices and hence are suitable for performing switching functions. The various electronic tubes, both vacuum and gas-filled, are essentially voltage-sensitive elements capable of being driven to a conducting or non-conducting state by control signals at a low energy level. Varistors and transistors, on the other hand, are primarily current-sensitive elements again functioning at low energy levels. Varistors offer either a low or a high resistance to current flow, depending upon voltage polarity, while transistors incorporate current amplification. Comparatively speaking, vacuum and gas tubes operate at higher voltage levels, and the semi-conductors at lower voltage levels. The chief advantages of electronic switching devices are high speed and sensitivity. Offsetting these factors, at least at the present time, are lower life expectancy and reliability, and probably less flexibility as circuit elements, as compared to relays. Also important ln a negative sense is the relatively heavy power drain required by hotcathode tubes. The design of electronic switching circuits is a young and dynamic art, subject to almost daily change and development, and advancing on many fronts. For this reason, no comprehensive treatment of the subject will be attempted in this volume. Rather, this chapter will discuss the basic characteristics of electronic apparatus that make it suitable for switching, together with some of the fundamental switching circuit arrangements. Examples of typical functional electronic circuits will be presented in later chapters. The theory of gas discharge, electron flow, and other physical phenomena related to the science of electron tubes and semi-conductors has been omitted since it is assumed that the reader is familiar with such concepts. It will be seen that electronic circuit design follows two main trends. One is to take advantage of specific characteristics of apparatus to perform particular switching functions. For example, gas tubes with multiple elements arranged in a certain geometrical relationship have been designed to count pulses. The other trend is to utilize electronic

- 202 -

Elementary

Electronic

Switching

203

Circuits

elements in the same sense that relays are used, that is, to interconnect the elements in patterns that satisfy the "and" - "or" logical requirements of the problem. Both these methods have validity and enhance the flexibility of tubes and semi-conductors as switching devices. 10.1

CHARACTERISTICS SEMI-CONDUCTORS

OF ELECTRON TUBES AND AS SWITCHING ELEMENTS

Hot-Cathode Vacuum Tubes. The construction and characteristics of the many hot-cathode vacuum tubes in general use are well known. The output of such tubes as used in switching applications usually consists of a voltage condition on either anode or cathode, or of a current flow in the cathode-anode circuit, with the tube in either case fully conducting or completely cutoff. This output depends upon potentials on one or more control grids which provide the means for setting up a particular output condition as a result of one or more input conditions. Two separate output leads, one from the anode and one from the cath(!de, can be furnished to external circuits; and, when the tube is nonconducting, the two output leads are isolated. The vacuum tube may be used for impedance transformation, as in the cathode follower circuit for which the input impedance is high and the output impedance is low. Vacuum tubes are extremely sensitive and very rapid in response to control signals, their speed being dependent upon the small interelectrode capacitances and the electron transit time from cathode to anode. Unfortunately for some possible switching applications, it is difficult to "lock operated" a single vacuum tube in the manner of a relay•. In general, the control grid potentials must be maintained even after the tube has been operated. Also, the vacuum tube has the disadvantage of requiring either standby filament current or sufficient time to heat the cathode adequately before the tube is called upon to operate. Hot-Cathode Gas Tubes. There are three types of hot-cathode gas tubes available for general use: the diode, the three-element thyratron, and the screen-grid (four-element) thyratron. The diode, normally employed as a non-linear unit capable of conduction in only one direction, rests in a non-conducting, or de-ionized, .state until the anode-to-cathode voltage is raised to approximately +15 volts. At this breakdown potential, the tube ionizes and remains ionized until the anode-to-cathode voltage is reduced below +15 volts. Thus, the sustaining voltage for the diode is approximately equal to the breakdown potential. In the ionized state, the voltage drop from anode to cathode of the tube is substantially constant over a wide range of current values. • A non-conducting operated.

tube may be referred

to as unoperated,

and a tube passing current

as

204

The Des~gn of Switching

Circuits

The three-element thyratron may be considered as a gas-tube diode with an added control grid. When the voltage on this control grid current conis sufficiently low with respect to the cathode, negligible duction occurs from anode to cathode. When, however, the grid voltage is raised to a poini more positive than the so-called cutoff value, the gas ionizes and conduction commences. After tube ionization, the.grid loses control over the discharge, and the anode-to-cathode voltage drop is maintained at a sustaining value of about +15 volts. The tube is extinguished by reducing the anode-to-cathode voltage to zero for sufficient time to de-ionize the gas. The grid cutoff value (the value of grid voltage just negative or that at which the tube is permitted to fire) depends upon the potential of the anode. An increase in anode voltage increases in a negative direction the voltage necessary to keep the tube cut orr. Two varieties of three-element hot-cathode gas tube are avail. able: the negative-grid and the positive-grid thyratrons. These designaof the tubes; the grid cutoff tions refer to the cutoff characteristics voltage for the negative-grid tube is of the order of two to ten volts tube is about nine negative, and the cutoff voltage for the positive-grid volts positive in a typical tube. The four-element thyratron is similar to the triode with the addition of a screen grid. This screen grid tends to increase the control grid impedance by reducing the current drawn by the control grid when that grid is near the cutoff point. In addition, the screen grid provides a means of control or the ionizing potential of the tube.

In accordance with the characteristics of gaseous discharge, large currents may be passed through the tube with low voltage drop. In this connection it should be noted that the life of a gas tube is a function or the time during which it conducts as well as of the magnitude of the conducted current. The useful life, then, may be increased by allowing only low values of current to be passed through the tube; or, if high current is necessary, by permitting it to now only momentarily• The hot-cathode gas tube, like the vacuum tube, requires standby filament current. on the The speed of response of the tube is contingent primarily ionization and de-ionization times of the tube. Depending upon the gas, to several the ionization time ranges from a fraction of a microsecond of the order of a microseconds; the de-ionization time is ordinarily hundred to a thousand microseconds, though lower values have been achieved. The tube, then, can respond very rapidly to input signals applied to operate the tube, but considerably more time must be allowed for extinguishing the tube. •

Elementary

Electronic

Switching

Circuits

205

I(

SA z STARTER SK : STARTER SG • SCREEN

Fig. 10-1

ANODE CATHODE GRID

Representative

A • MAIN ANODE K z CATHODE

Cold-Cathode

Gas Tubes

Cold-Cathode Gas Tubes. At present, the cold-cathode tube is the most attractive type of electron tube from the point of view of switching applications where extreme speed is not necessary. Although twoelement and three-element cold-cathode tubes are most common, many other types with a greater number of elements have been constructed for special purposes. Certain of these more intricate tubes are illustrated in schematic form, together with the cold-cathode triode, in Fig. 10-1. Those tubes with multiple main anodes provide for several outputs which are mutually isolated until the tubes fire; those with multiple starter anodes allow the tubes to be ionized by any one of several isolated input conditions. Although the theory of gaseous discharge in the cold-cathode tube is fundamentally different from that in the hot-cathode gas tube, the two types of tube are somewhat similar in operating characteristics. As in the hot-cathode tube, sufficient potential difference between two electrodes of the cold-cathode tube causes the tube to ionize and the anodeto-cathode gap to conduct. A tube remains ionized until the potential from anode to cathode is reduced below the sustaining value. The firing and sustaining voltage values for the cold-cathode tubes differ from those encountered in the hot-cathode gas tubes. A typical cold-cathode diode ionizes at an anode-to-cathode potential of about +70 volts, and remains ionized until this potential is reduced below +60 volts. The three-element cold-cathode tube employs a starter

206

The Design of Switching

Circuits

anode (rather than a control grid as in the corresponding hot-cathode tube) which must be raised to a potential of the order of about +70 volts to fire the tube. In firing this tube, the starter-anode-to-cathode gap is ionized, and the discharge is transferred to the main-anode-to-cathode gap. To effect this transfer, the main anode potential must be above the sustaining value of the main gap but not above the self-ionizing value. The sustaining potential of the main gap is normally of the order of 75 volts, and the self-ionizing potential varies from about 150to 250 volts. After the main-anode-to-cathode gap is ionized, the starter anode has no further effect on conduction until the tube is extinguished. Such properties as high sensitivity, low conduction impedance, multiple outputs, and multi-control pertain to the cold-cathode tube. Moreover, since the cold-cathode tube has no filament, no standby current is consumed. The speed of response, though somewhat less than that of the hot-cathode gas tube, is sufficient for most applications. The ionization time depends upon the time necessary to transfer the discharge from the starter gap to the main gap, and it is generally less than a hundred microseconds. Main gap de-ionization times are of the order of one to ten milliseconds.. Because of its suitability to switching circuits, the electron tube circuit examples contained in the remainder of the chapter are, in the majority of cases, based on the cold cathode-tube. Semi-Conductors. Although the advent of semi-conductors into the field of switching circuits is relatively recent, two types of semiconductors - the non-symmetrical germanium or silicon varistor, and the germanium transistor - appear to be suitable as switching devices in numerous applications. As with the cold-cathode gas tube, no standby filament power is expended and no warm-up time is required before operation. Upper-frequency limits of operation lie above the megacycle mark; the germanium varistor retains its non-linear characteristics up to a frequency of the order of 100 to 1000 megacycles, and the maximum useful frequency for the transistor appears, at present writing, to be well above 10 megacycles. The non-symmetrical varistor is a non-linear element with a high ratio of non-conduction ("reverse") resistance to conduction ("forward") resistance, the actual values of these resistances depending upon the voltages applied to the element. A typical varistor may have a reverse resistance of 50,000 ohms at 50 volts across the unit, and a forward resistance of 300 ohms at 1 volt. The germanium transistor is an amplifier element equipped with three electrodes - the emitter, the collector, and the base - as shown in Fig. 10-2. In normal operation, the emitter voltage Ve and current le are positive, and the collector voltage Ve and currentlc are negative•

Elementary

Electronic

Switching Circuits

:t•

EMITTER

--..-

Vel

207

COLLECTOR

le

fv,

Fig. 10-2 Symbolic Diagram of Transistor

As a result of these polarities, the emitter-to-base circuit acts as a semi-conductor rectifier conducting in the forward (low-resistance) direction, and the collector-to-base circuit acts as a rectifier conducting in the reverse (high-resistance) direction. However, there is a large degree of interaction between these two circuits: a change in the magnitude of Ve or le causes a corresponding change in the magnitudes of Vc and le. Ordinarily, the emitter is used as the control electrode, the input signal to the transistor varying the emitter voltage or current. Where the transistor is used primarily as an amplifier, the voltage or current output is obtained from the collector electrode, although for certain applications the output may be derived from the emitter orbase terminals. The transistor may be employed to take advantage of its nonlinear properties in addition to its use an an amplifying device. Unlike the vacuum tube, the transistor cannot be placed in ·a cutoff state by a large negative potential on its control electrode. However, when such a potential is applied, the amplification factor of the transistor is a small fraction of unity; for a less negative control condition, the voltage gain is appreciable•. Also in contrast to the vacuum tube, the input (emitter-tobase) impedance of the transistor (as normally used) is low, and the output (collector-to-base) impedance is high; typical values of input impedance and output impedance are200 ohms and 10,000ohms, respectively.

10.2 BASIC TUBE AND SEMI-CONDUCTOR CIRCUITS FOR SWITCHING APPLICATIONS Lock-In or Memory Circuits. In most switching circuits, certain relays are required to operate in response to some particular condition and to remain operated, or locked-in, until a second particular condition obtains. As discussed in previous chapters, such a requirement is met by a relay furnished with a holding path, the relay operating when its operating path is closed, and locking operated until the holding path • In actuality,

the transistor ls a current amplifier with a current gain. Specific values are not quoted here since the device Is still in the comparatively early developmental stage.

The Design of Switching Circuits

208

is opened. The relay, by virtue of its associated control networks, has two stable states: when it is unoperated, it remains unoperated until the operate condition is applied; and when it is operated, it remains operated until the release condition is applied. Elements with a similar characteristic are necessary in many electronic switching circuits. Hot-cathode or cold-cathode gas. tubes may be used directly as lock-in elements since, after being ionized,

J. .I.

TIME

Ep

+

TIME

OPERATION OF COLD· CATHODE DIODE

A

+ 1-IM

zm

iu::::I

11:111::::, ::Ill:

u:z: I-

POSITIVE DYNAMIC RESISTANCE

1'--4--

I

NEGATIVE DYNAMIC RESISTANCE DYNAMIC RESISTANCE

0 ANODE-CATHODE POTENTIAL

+

60V. 70V.

CHARACTERISTIC CURVE FOR COLD· CATHODE DIODE

B Fig. 10-3 Operating Characteristics

of Cold-Cathode

Diode

Elementary

Electronic

Switching

Circuits

209

they remain in a conducting state until the potential between conducting electrodes is reduced below the sustaining value. Inherently, then, gas tubes are devices with two stable states. As an example of an elementary gas-tube lock-in circuit, consider the cold-cathode diode ci:r:cuit shown in Fig. 10-3A. This circuit consists of the tube in series with a load resistance in the cathode circuit and a variable potential Ep applied to the anode. As the anode potential is increased from zero in a positive direction, the voltage drop Ek across the load resistance remains at zero as long as Ep remains below the firing potential of the tube, about 70 volts for a representative cold-cathode tube. When Ep reaches this ionizing value, the tube conducts; and, as the characteristic curve in Fig. 10-3B shows, the voltage drop across the tube is reduced to 60 volts. Since at this as Ep is inmoment Ep equals +70 volts, Ek is +10 volts. Thereafter, creased from +70 volts, Ek increases by alike amount, the voltage drop across the tube remaining almost constant. Note that the characteristic curve of Fig. 10-3B consists, in effect, of two positive dynamic resistance regions - one of very low current (very high resistance) and the other of moderately high current (low resistance) - separated by a region of negative dynamic resistance. The tube, in firing, passes through this negative resistance region between non -conduction and conduction. When Ep is reduced, in the circuit of Fig. 10-3A; the current through the tube diminishes until, at a very low value of current, the knee ("a" in the diagram of Fig. 10-3B) of the characteristic curve is reached, the voltage across the tube begins to decrease rapidly, and the tube de-ionizes. This occurs at a value of Ep of slightly less than +60 volts. As a lock-in element, then, the anode may be biased at a point between the sustaining and firing voltages. The tube is operated by application of a pulse sufficient to raise the anode-to-cathode potential to the ionizing value, and is released by reduction of the anode-tocathode potential to a point below the sustaining value. Release can also, of course, be obtained by opening the conduction path with a switch Since the margin between ionizing and sustaining or it~ equivalent. values may be small (in this case, 10 volts), a diode is not practical as a lock-in element in some applications. The gas-tube diode as a lock-in device also has the disadvantage that the input and output leads are not isolated during the interval of the gas-tube tube conduction. If isolation at all times is necessary, triode may be employed, as shown in Fig. 10-4. In this circuit, the main anode is connected to a positive voltage source, and the cathode is connected to ground through a load resistance and a contact or some other means capable of opening the main-anode-to-cathode conducting

The Design of Switching Circuits

210

(ST)

1 -

l

=

-=-

-=Fig. 10-4 Operation of Three-Element

l ••

Gas Tube

path. The main-anode voltage supply is of a value lower than that necessary for ionizing the main conducting gap but greater than the maingap sustaining potential. With the circuit at rest, the starter anode is at ground potential. When a sufficiently positive voltage Est is applied to the starter anode, in this instance by the operation of the (ST) key, the starter-anode-tocathode gap is ionized, the discharge is rapidly transferred to the main anode, and the main conducting gap is ionized. After the main gap is fired, the starter anode exercises no further control until the tube is de-ionized. Main-gap de-ionization is accomplished by reducing the main-anode-to-cathode potential below the sustaining value. This is done in the circuit of Fig. 10-4 by opening the conduction path through the cathode circuit. The operating input condition for the cold-cathode tube circuit consists of a positive voltage signal, and the corresponding output consists either of a current flow through the load resistance or a potential Ek measured across this resistance•. Since the starter gap requires only a few micro-amperes for ionization, the source of the input signal may be of high impedance. The transistor as a lock-in element is somewhat similar to the cold-cathode diode. This similarity is clearly indicated, for example, by the curve of emitter current as a function of emitter voltage shown in Fig. 10-5. This characteristic curve applies to the circuit of the same figure where the collector voltage is held constant and the resistan~e Rb in series with the base is large. In this figure, the arrows associated with le, le, and Ve indicate the sense of positive current flow and positive potential. Like the curve representing the characteristics of the cold-cathode diode, the curve of Fig. 10-5 indicates two regions of positive dynamic resistance separated by a region of negative resistance. Therefore, the transistor connected as shown in the • In some gas-tube circuits, where the starter-anode-to-main-anode ciently high, Ionization between these two electrodes may occur current may be used as an output indication.

potential and the

is suffiresulting

Elementary

Electronic

Switching Circuits

211

figure has two stable states, one characterized by low negative values of emitter current le, and the other characterized by high positive values of le. By selecting appropriate values of emitter bias E and the series resistance Re, the circuit can be pulsed from one stable state to the other, thus fulfilling the general requirements of a lock-in device. The collector-to-ground characteristic of the transistor with constant emitter potential and a high resistance in the base circuit also indicates the presence of two stable states. Here, however, the collector current le is negative in both states, the difference between states being one of current magnitude. +Ve

le

COLLECTOR

NEGATIVE DYNAMIC ESISTANCE

\ POSITIVE DYNAMIC RESISTANCE

I

POSITIVE DYNAMIC RESISTANCE

-Ve

Fig. 10-5 The Transistor

Used as a Lock-In Device

Lock-in circuits may also be composed of electronic elements which do not, in themselves, exhibit lock-1.n characteristics. Consider, for example, the vacuum triode. Two triodes may be interconnected in such a manner that when one is conducting, the other is cut off. With appropriate values of components such an arrangement has two stable states, the circuit remaining in one until forced into the other by external means. One triode circuit of this type is the Eccles-Jordan circuit, shown in somewhat modified form in Fig. 10-6. Assuming that, with the circuit at rest, tube T 1 is conducting and tube T 2 is non-conducting, a negative pulse applied to lead 1 decreases the grid potential of T 1 . The plate potential of T 1 rises sharply, carrying with it the grid potential of T2 , thus allowing tube T 2 to conduct. The accompanying decrease in plate potential of T 2 forces the grid potential of tube T 1 below cut off. Tube T 2 remains in a conducting state and T 1 remains in a non-conducting state, until a negative pulse applied to lead 2 returns the circuit to its original state. Thus, the presence of a negative pulse.on lead 1 is the operating condition. An indication of the state of the circuit may be obtained by examining the plate voltage or plate current of either

The Design of Switching Ci-rcuits

212

tube. In Fig. 10-6, the plate current of tube T 2 gives an output voltage indication. In applications of this circuit, it is necessary to insure that the desired tube is conducting (in this case T 1 ) before the circuit is required to function. By suitable changes in the arrangement of Fig. 10-6, the circuit may be made to operate and release on consecutive re-appearances of the same input condition; i.e., the negative pulse. A circuit operating in this manner is evidently a pulse-frequency divider, since it provides a given output indication for every alternate input pulse, and corresponds in action to the relay pulse-frequency divider presented in Chapter 8.

OUTPUT

-v-

2

!._ ___ J

•-y-

---,t--+----➔

INPUT

lNPUT

-=

Fig. 10-6 Eccles-Jordan

Circuit as a Lock-In Element

Lockout Characteristics. A characteristic of gas tubes which is the lockout, of great importance to one functional circuit in particular, which is discussed in a later chapter, concerns the operation of tubes connected in parallel. It has been experimentally determined that if two or more gas tubes are provided with a common load resistance, as in Fig. 10-7, either in the anode or in the cathode circuit, only one tube, except in rare cases, will ionize and remain ionized even if firing potentials are applied to the several tubes either simultaneously or in sequence. This phenomenon is due to the region of negative resistance in the characteristics of the gas tube through which the tube passes in the range

Elementary

Electronic

Switching

Circuits

213

J_

J

R

Fig. 10-7 Two Gas Tubes with a Common Load Impedance

between breakdown and the stable sustaining voltage. In this region, as the current through the tube increases, the voltage across the b.lbe decreases, tending to prevent other tubes with the common load resistance from firing .. However, if two tubes reach the breakdown potential simultaneously and then travel through this unstable region exactly together, it is possible for the two tubes to remain ionized, since in the stable condition of normal conduction the characteristics exhibit a positive resistance. To reduce this slight possibility of dual operation,' the common load resistor is replaced by an impedance including an inductive element. The effect of the inductance is to increase the time interval necessary for the tubes to traverse the unstable negative resistance re gion, decreasing the likelihood of the two tubes passing through this

1 -I-

Fig. 10-8

=

A Modlflcatlon or the Circuit or Fig. 10-7

The Design of Switching Circuits

214

region exactly together. A possible second modification is that of employing multi-element tubes which ionize in successive stages, corresponding stages having common inductive impedances. One such arrangement is shown in Fig. 10-8. The tubes in this circuit have a separate starter cathode, with the result that the startergap firing circuit is isolated from the main anode-cathode gap. The tubes are fired in two stages: first, the starter gap is ionized, and then the discharge transfers to the main gap. In the circuit of Fig. 10-8, the two tubes can be fired coincidently only if both pass together through the unstable region for each ionization: that of the starter gap, and that one such of the main gap. If the probability of two tubes traversing region is very small, the probability of covering all such regions together is infinitesimal. SIGNAL

INPUT

EsI:.A..A...fL..

OUTPUT

0

E

_j\]_Eo

GATE GATE OPEN ----._ GATE CLOSED -

_J

r-7""i Li Eg

Fig. 10-9 Elementary Diode Gate Circuit

Gate Circuits. A basic element which is necessary in many switching circuits is one which is capable of opening and closing a signaling or control path between other elements. Such an element is, in effect, a switch which is either open or closed to signals passing along a lead, depending upon the conditions controlling the switch. In usual terminology, this electronic switch element is called a "gate". A gate circuit may operate in one of a number of different ways. It may act to alter the impedance of a path carrying signals through the gate, or it may act as an amplifier which is characterized by an approximately zero gain when the gate is closed and appreciable gain when the gate is open. In most cases, a gate circuit is designed to operate on a path carrying signals comprising positive or negative pulses rather than on a path carrying a-c signals. Many gates have non-linear characteristics which make them unsuitable for a-c transmission.

An elementary gate circuit employing a vacuum diode is illustrated in Fig. 10-9. In this circuit, the bias voltage E is maintained on the output lead as long as the anode-to-cathode potential is negative. The magnitudes of the signal pulse E8 and the gate-open voltage condition Eg are such that the anode-to-ground potential remains less than

Elementary Electronic

Switching Circuits

215

the cathode-to-ground potential unless the signal pulse and the gateopen voltage condition are present simultaneously. When both signal pulse and gate-open voltages are applied to the anode, the tube conducts, raising the cathode or output voltage. The wave shape of the output pulse is similar to that portion of the signal pulse which occurs during the period of tube conduction. There is an inherent loss in signal amplitude between input and output due to the voltage-dividing action of the circuit resistances. A varistor rectifier may replace the diode in the circuit of Fig. 10-9 if the specific requirements for the circuit permit. Comparison of the characteristic curves for the vacuum diode and the varistor, shown in Fig. 10-10, indicates that the forward resistance of the varistor is less than the conducting resistance of the diode, and the reverse resistance is less than the diode non-conducting resistance. Therefore, the output voltage level may be affected by signal pulses with the gate closed if a varistor is substituted for the diode, although the conducting loss through the varistor when the gate is open will be less. Where gates are to be constructed using varistors, slightly more complex circuit arrangements are often employed to increase the gating efficiency. Consider,for example, the gate of Fig. 10-11. The operation of this circuit in either the gate-open or gate-closed condition is to some extent equivalent to that of a transmission filter T-section. When the gate voltage is negative, series varistors V 1 and v2 are biased in the reverse direction, and the shunt varistor v3 ls biased in the forward direction; when the gate voltage is positive, the directions of the varistor biases are reversed. Inspection of the varistor characteristic curve of Fig. 10-10 shows that the small-signal (small in comparison to the magnitude of d-c voltage acroff's the varistor) a-c resistance is

+ Iy

+Ev

-Ey

-Ip VOLTAGE-CURRENT CHARACTERISTICS OF VACUUM DIODE

Fig. 10-10 Characteristics

-Iv VOLTAGE-CURRENT CHARACTERISTICS OF NON· SYMMETRICAL VARISTOR

of Vacuum Diode and Varistor Rectifiers

The Design of Switching Circuits

216

high when d-c current flows through the varistor in the reverse direction, and low for d-c currentin the forward direction•. Therefore, for a negative gate voltage the series resistance is high to the signal, and the shunt resistance is low. For a positive gate voltage, the converse is true: the series resistance is low and the shunt resistance is high. The resistor R is placed across varistor v3 to increase the forward current through V1 and v2 , thus effecting a greater reduction in series resistance when the gate is open. Vz

SIGNAL INPUT LOW

OUTPUT

IMPEDANCE SOURCE

v.,

R

GATE

+,-,

~_J~ Fig. 10-11 Varistor Gate Circuit

A simple gate circuit based on d-c control of small-signal a-c varistor resistance is shown in Fig. 10-12. As in the circuit just discussed, a negative gate voltage increases the series resistance to the signal voltage while the series resistance is reduced by application of a positive gate voltage. No shunt varistor appears in the arrangement of Fig. 10-12, although more complicated varistor networks making use of variable resistance shunt elements may be constructed.

If signal gain is desired when the gate is open, a gate employing an amplifying tube may be used. One such gate circuit is shown in Fig. 10-13. In this arrangement, the grid of the triode is held below cutoff except during the interval when the gate is to be open. During this interval, the gate voltage biases the grid approximately to the cutoff point. Thus, positive signal pulses placed on the grid have no effect on the output voltage except while the gate is open. The magnitude of the signal voltage must be limited to a value insufficient to raise the grid above cutoff when the gate-open voltage is not present. A transistor may also be used for gating purposes where signal gain is desirable. The operation of a transistor in a gate circuit, as for example in the circuit of Fig. 10-14, is very similar to that of a vacuum triode. With the gate voltage sufficiently negative, the input signal • Note that the small-signal a-c resistance of the varistor at a given d-c current level is given by the inverse slope of the characteristic curve at that current: Ra-c = d(Ev)/dUv); while the d-c resistance at the same current Is the ratio of values of Ev and Iv coordinates: Rd-c "' Ev /Iv•

Elementary

Electronic

Switching Circuits

217

[=u•

SI~ IN~

GATE

+c:::J ~_J

L

Fig. 10-12 Gate Circuit Arranged for Voice-Frequency

Signals

causes only a very small change in collector current; with the gate voltage at ground potential, the input signal is amplified by the transistor action and appears on the output lead. "And"- "Or" Circuits. In most relay circuits, the state, at a given instant, of a particular relay (operated or unoperated) or lead (activated or not-activated) depends upon the combination of states of other relays or leads at that moment. This is also true of electronic control circuits in that the state of some one element depends upon the state of a number of other elements. The design of an arrangement which will exhibit this required dependency among elements involves the use of circuits which associate the independent elements with the dependent elements. These circuits correspond, to some extent, to the contact networks constructed on a group of relays, allowing tliese relays to control the action of certain other relays. There are two basic forms of these electronic control circuits: the "and" circuit and the "or" circuit. In both forms, a number of input

OUTPUT

v

SIGNAL

INPUT

__A_A_A_

GATE

----0 ILCUTOFF

Fig. 10-13 Elementary

Triode Gate Circuit

The Design of Switching Circuits

218

OUTPUT

SIGNAL INPUT

__A_J\_A_

,1

GATE

o_J

l

I-=-

L

Fig. 10-14 Transistor

Gate Circuit

terminals and a single output terminal are provided. For the "and"circuit, the output terminal is activated only if all input terminals are activated; for the "or" circuit the output terminal is activated if any input terminal is activated. Therefore, the "and" circuit may be considered as corresponding to a relay circuit in which each input lead is connected to a relay winding, and the output lead is controlled by a series chain of make-contacts placed on the relays. Similarly, the "or" circuit corresponds to a relay circuit in which the output lead is controlled by make-contacts placed in parallel. An example of an "and" circuit employing either vacuum diodes or non-symmetrical varistors is shown in Fig. 10-15A. In this circuit, the rectifier units are poled in such a manner that the output lead can E +E

,---1--i

R

on_ INPUT LEADS

2

on

3

oTl_

OUTPUT

FL

A I INPUT LEADS

:

~

I

•

LEAD

OUTPUT LEAD 0

B Fig. 10-15 Two Forms of "And" Circuits

E 0

Elementary

Electronic

Switching Circuits

219

never assume a positive potential appreciably above that of the least positive input lead.Assumingthat the input leads are at zero volts when unactivated, and that the resistance R indicated in Fig. 10-15A is large in comparison to the forward resistance of the rectifier units, the potential of the output lead remains close to zero volts as long as any of the inputs are unactivated. The input leads, in this case, are activated by signals of +E volts, and when all these leads are thus activated, the output lead is allowed to rise to a point somewhat less than +E volts. The exact positive potential which the output lead may reach depends upon the comparative values of resistance R, the external resistance faced by the output lead, and the back resistance of the rectifiers. In the circuit of Fig. 10-15A, power to the output circuit is supplied through resistor R by the bias battery E. If it is permissible to draw power from one of the circuits controlling an input lead, the circuit of Fig. 10-15B may be employed. Inspection shows the operation of this circuit to be similar to that of Fig. 10-15A with the distinction that the bias source of constant potential +E is replaced by an input signal source which has a potential of either zero volts or +E volts. Here, the resistance of R must be considerably less than the reverse resistance of a rectifier unit in order to maintain the potential of the output lead near zero volts when input lead 3, in the figure, is at zero volts, and leads 1 and 2 are at +E volts. Conversely, R must be substantially greater than the forward resistance of the rectifiers to reduce the voltage rise on the output lead when terminal 3 is at +E volts and terminals 1 and 2 are at zero volts. The corresponding "or" circuit, as shown in Fig. 10-16, is equally simple. As a result of the polarity of the rectifier elements, the potential of the output lead is close to +E volts when a potential of +E volts is applied to one or more of the input leads. Again, the actual voltage reached by the activated output lead depends upon the resistance R and also upon the forward resistance of the rectifiers. In this circuit, power to the output circuit is supplied by whichever inputs are activated. The principal effect of the rectifiers is to isolate the input leads. OUTPUT LEAD INPUT

o...rl...

LEADS

R

Fig. 10-16 An "Or" Circuit

220

The Design of Switching Circuits

r-----,.---~------....-

- -+---i1_

R

-:-

-:-

-:-

IA)

IB)

~

f~I•~

_______

p

IC)

..

I

ID)

mt

1·~~-~

Fig. 10-17 Gas-Tube Application or "And" Circuit and Equivalent

l•q_

1

___

i

Relay Circuit

As an elementary example of the utility of these circuits, consider the circuit of Fig. 10-17 in which the cold-cathode gas tube D is to be fired when all three control tubes A, B, and C are in a state of conduction. The cathodes of the control tubes are connected to an "and" circuit similar to that illustrated in Fig. 10-15; the output of this "and" circuit - point Pin the diagram of Fig. 10-17 - is joined to the starter anode of tube D. As a result of the characteristics of the "and" circuit, point P maintains a potential nearly equal to that of the least positive cathode, assuming the cathode resistances to ground to be small. This potential is insufficient to ionize the starter gap of tube D when any of the tubes A, B, and C are non-conducting. As each control tube fires, its cathode voltage rises; and when the last of the three tubes fires, point P and the starter anode of tube D take on a potential approximately equal to the cathode voltage of the ionized tubes. Thus, the starter gap of tube D ionizes and the rube fires. All four tubes remain ionized until the connection between the main anodes and the voltage supply is opened. Note the equivalence in action of this circuit and the relay circuit also shown in Fig. 10-17. Similar circuits can easily be visualized in which a varistor "or" circuit permits a tube D to ionize if any of tubes A, B, or C are conducting, or prevents a tube D from ionizing if any of tubes A, B, or C are conducting, etc.

Elementary

Electronic

Switching Circuits

221

The switching algebra, introduced in a preceding chapter, may be used as an aid to the combination of "and" and "or" circuits to fulfill stated requirements for activating a particular lead. Here, the algebraic expression which is equivalent to a statement of requirements is written in terms of leads .activated and unactivated rather than in terms of relays operated and unoperated. For example, if lead D is to be activated when lead B and either lead A or lead C are activated, the corresponding algebraic expression is: f(D) = B + AC. Except for this difference in reference, the algebra as applied to the activation of leads ,-e+cAc)-1

oIL

I

eo---------------'VVv---,

I I

Rz V3

.....---1--i--

...

L _____

oll

co---------

I

----4~---+----00 I "AND" CIRCUIT I

A0-----+-➔,._-

_j

..

I I I I I I L

"OR" CIRC~IT _____

_j

I

,------, A+

I

B

A

I

I

I

I

I

L "AND" _____CIRCUIT Be>---

,------7 I

B +C

I

I _J

I I

I

.....,:-1-- .... - ...... I I

I C

l(A+BllB+c)7

t-1-+---t~-➔-----00

I CIRCUIT J L "AND" ______ B

I II

1 ''0R"c1R;u1T L _____

_J

Fig. 10-18 Two Combinations of "And" and "Or" Circuits Corresponding to the Expression: f(D) = B + AC

222

The Design of Switching Circuits

is identical to that used in the design of relay circuits, and may be similarly employed in the manipulation and simplication of circuit configurations. The "and" and "or" circuit combination which corresponds to an algebraic expression may be derived by inspection. Any such expression consists of variables which are associated by addition ("and"} and multiplication ("or"). Thus, all that is necessary in obtaining a circuit equivalent to the algebraic expression is to separate the expression into "and" and "or" associated terms, and then to draw the corresponding circuit, representing the "and" associations by "and" rectifier circuits and the "or" associations by "or" rectifier circuits. As an illustration of this method, consider the example mentioned above: f(D) = B + AC. Here, the expression indicates that leads A and C should be associated by an "or" circuit, and the result (the output lead of the "or" circuit) should be associated with lead B by an "and" circuit. This arrangement is shown in Fig. 10-18A. Study of the figure indicates that the original requirements have been fulfilled; lead D is driven positive for a P.OSitivepulse on lead B occurring simultaneously with a positive pulse on lead A or lead C. However, it is clear from an inspection of Fig. 10-18A that the values of resistances R 1 and R2 are somewhat critical. As stated in connection with the discussion of the basic "and" and "or" circuits, resistances R 1 and R2 should be greater than the forward resistance of a rectifier element, and R 2 must be less than the reverse resistance of a rectifier element and also less than the output resistance faced by lead D. Moreover, R 2 must be greater than R1 in order to minimize the potential of lead D when lead B is activated and leads A and C are at zero potential. Also, when "and" and "or" circuits are combined in this manner, care must be exercised to maintain low resistances at the sources of the input signals. If the source resistance at terminal B, for example, is high, positive voltag·es on terminals A and C may cause an appreciable positive potential on output terminal D, giving a false output indication. Analysis of this kind must be employed in the determination of the values to be assigned to the resistances employed in combinations of "and" and "or" circuits. Evidently, as the number of circuits which are to be combined increases, it becomes more difficult to obtain sufficient margins for practical application. In some cases it may be necessary to regenerate the combination of activated and unactivated conditions on an intermediate output lead by a gas-tube or vacuum-tube circuit before allowing these conditions to affect further circuit action. In the circuit of Fig. 10-18A, power is supplied by the source connected to lead B to the external circuit connected to lead D. Under

Elementary

Electronic

Switching Circuits

223

certain conditions it may be desirable to distribute the power load over the input sources in some other manner. To effect this distribution, the circuit may be modified by such methods as interchanging the resistor and rectifiers in the legs of the "and" circuits - in this case interchanging rectifier v3 and resistor R 2 . Another method is to manipulate the original algebraic expression from which the circuit configuration was derived. As an example, the original expression f(D) = B + AC is equal to: f(D) = (B + A)(B + C). This latter form suggests a combination of two "and" circuits associating leads B and A and leads B and C, respectively, with their outputs associated in an "or" circuit, as shown • in Fig. 10-18B. In this arrangement, power is supplied by leads B and C, a result which could not have been obtained as easily by direct manipulation of the circuit of Fig. 10-18A. A

A

+

C

B

AB=C

A+B=C

Fig. 10-19

~ ~· -

IA' B',.C'l

(A'+B'=C'l

A

B of Diode "And" and "Or" Circuits

Symbolic Representation

A convenient system of symbolizing these "and" and "or" circuits is indicated in Fig. 10-19. Here, the basic "and" or "or" element is provided with multiple input leads and one output lead. The characteristic behavior of the element is specified by the polarity signs adjacent lo the input and output leads. For example, the symbol of Fig. 10-19A indicates that when A is positive(+) and B is positive(+), C is positive (+); the symbol of Fig. 10-19B indicates that when A is positive Q!: B is positive, C is positive. Thus, these symbols correspond to the circuit forms shown in Figs. 10-15 and 10-16. The actions of these two symbolic circuit elements may be expressed in another way. In the first, that of Fig. 10-19A, C is negative (-), with respect to its activated condition, when either A or B is negative. For the element of Fig. 10-19B, C is negative when both A and B are negative. Now if, in an algebraic expression, the prime(') is used to indicate the negative (-) condition and the absence of the prime indicates the positive (+) condition, the behavior of the two elements may be expressed in either of two forms: For Fig. 10-19A, For Fig. 10-19B,

A+ B = C or A'B' = C' AB=C or A' + B' = C'

Algebraically, the two forms characterizing the action for each element are equivalent. It is interesting to note that the algebraic expression

The Design of Switching Circuits

224

A

A

+

C

C

8

8

AB'=

A+B'=-C ( A' 8 =-C'l

(A'+B

C

= C 1l

A

C

8 A'+B'=-C

A' 8 1 =-C lA+B =-C1l

(A 8 =-C'l

Fig. 10-20 Additional "And" and "Or" Elements

indicates that the "and" element of Fig. 10-19A can be considered an "or" circuit for negative activation of its leads. Similarly, the "or" element of Fig. 10-19B is an "and" circuit for negative activation. In addition to these two symbolic elements, four other elements shown in Fig. 10-20, are logically possible. Each of these is shown with its characteristic algebraic expressions. The physical realization of these four elements necessitates some means for signal inversion through the use of vacuum tubes or transformers. As an illustration, a circuit corresponding to the symbolic element A + B : C' (or A'B' : C) is indicated in Fig. 10-21A. In this circuit, only when both grids A and B are raised in a positive (+) sense above cutoff does the anode C become negative(-) with respect to the anode supply voltage. Similarily, in a circuit for AB: C' (or A' + B' : C) in Fig. 10-21B, the anode C falls below the supply potential when either A or B becomes positive with respect to cutoff. The remainder of the symbolic elements may be realized with somewhat similar circuit arrangements. In fact any given gate, say the pentode of Fig. 10-21A, can be used to realize any of the six circuits by inserting signal-inverting elements (triode amplifiers or transformers) in the appropriate input and output leads.

A

= C'

A+B A1 8 1

OR

=C

A'+ 8 1

A

Fig. 10-21 Typical Elementary

= C'

A 8

A

OR

8

•

B "And" and "Or" Circuits

Providing

Signal Inversion

C

Elementary

Electronic

Switching Circuits

225

+

Fo--------------------.. Do+______

+, ORI-+ __

..+___ ,.)(

_

+

Eo-------"".

+

C ~-------~ +

Ao----1

,._____,.,. + (A+B'l

+ Bo-----

A

X= [-----~

SE0UENCEyc1RCUIT

CIRCUIT

2

CIRCUIT 3

A CIRCUIT I

CONTROL

y

CIRCUIT

2

CIRCUIT 3

y ~-------~-------SEQUENCEYCIRCUIT

COMMON CIRCUIT

B

CONTROL

y

SEQUENCE YCIRCUIT

I

I

CIRCUIT I

CIRCUIT

2

'(

CIRCUIT 3

C Flg. 11-23 Arrangements Involving Sequence Circuits

Circuits for Counting

269 Al

PERIOOS OF PATH CLOSURE OPEN INTERVALS

Bl

A2

92

A3

93

A4

94

7 77 7___ _

Fig. 11-24 Operating Diagram for Sequence Circuit

Although several of the counting circuits discussed earlier in this chapter are applicable in this example, the circuit of Fig. 11-10 is frequently used for requirements similar to those above. This arrangement, as employed here, consists of one two-relay prime counter per individual circuit, the prime relays operating immediately following control pulses and the nonprime relays operating during control pulses. If the path from the common circuit to each individual circuit is to be closed during the respective A- interval and may also be closed during the preceding open interval, a logical contact network is a transfer chain on the prime relays with outputs from the break-contacts, as shown in Fig. 11-25 (Network A). This has the advantage that the path to any individual circuit is prepared in advance of the interval in which it is needed. If circuit requirements permit, however, the common-toindividual circuit network could be placed on the nonprime relays, as shown for Network B in Fig. 11-25. In this case, closure of each succeeding path is delayed beyond the start of the corresponding A- interval by the operating time of the nonprime relay. After the final connecting circuit is served, it is necessary for the common circuit to open the path holding the counters operated, restoring the circuit to normal. This particular sequence circuit can be extended indefinitely by adding additional prime counters. By proper choice of control conditions, it can be made to fit many of the situations requiring sequential operation. Other forms of prime counter circuits can also be adapted to fulfill particular sequence requirements. A desirable objective in circuits to perform the sequence function is a circuit which requires but one relay per stage. This has been approached in the counting circuits discussed earlier in this chapter, but always at the additional expense of a group of common relays. A true relay-per-stage circuit can be designed, but is not often used for counting since it requires a heavy spring load on each relay for control and, to be completely reliable, may require two input pulsing leads.

The requirements for the circuit will be the same as those established for the previous sequence circuit with the additional provision that it be able to recycle, or act like a ring circuit. The attempt will be made to drive the circuit with a single pulse lead. A suitable operating plan to meet these requirements is shown on Fig. 11-26. Analysis of this plan indicates that the conditions for operating a typical relay are

270

The Design of Switching

Circuits

that the pulse be present, the preceding relay be operated and the second preceding relay be released. The relay must hold to off -normal ground during the interval following the end of the pulse and to the pulse ground during the next succeeding interval. A circuit designed to meet these conditions is shown on Fig. 11-27. The operating ground from the (CON) relay is carried to the winding of each relay through the break contact of a continuity-transfer in order to prevent a back-up of locking ground onto the lower branch of the pulse lead. The latter is used to hold the preceding relay during the pulse. An alternate method of preventing this back-up is a secondary A2

Al

A3

RELAY IA)---------------RELAY IA') RELAY 181 RELAY le') RELAY IC) RELAY IC')

A TO CIRCUIT I COMMON CIRCUIT -~1----r.----"

:

TO CIRCUIT 2

______,

TO CIRCUIT 3

-------NETWORK A

.__

------

L. _

__r------0-------

~

D

CONTROL LEAD

O.N.

-!----------.--+-+------1----1------+-+-------

NETWORK

r

:

r--------------------------------, ei: ------4--' ______ _:--o---l L.

---1

------------

TO CIRCUIT I

------------

TO . CIRCUIT 2

_____

TO CIRCUIT

3

B Flg. 11-25 Sequence Clrcult Based on Prlme Counters

J

COMMON CIRCUIT

Circuits

for Counting

271 Al

RELAY ICON) RELAY

(Al

RELAY (Bl RELAY

(Cl

RELAY

(0)

RELAY

Fig. 11-26

IE)

~

I

I

A2

A:5

i----,

t---1

I I I

I I I I

I I

I I

I I I I

I

A4

I I I I

I

~ :

I

I

I

I I

I

I

I I

I

I

A5

I

I

I

I

Plan of Operation for Sequence Circuit Employing One Relay per Stage

winding per relay for holding. If a pair of input leads, p 1 and p2 , are practicable, neither of these methods is necessary. The locking path for each relay is taken through a continuitytransfer on the succeeding relay in order to prevent momentary opening of the locking path when shifting from off -normal ground to pulse ground at the time this next relay operates. However, use of this continuitytransfer introduces an operating hazard in that the release of a relay momentarily connects ground back on the pulse lead during the bunching time of the continuity. If there is unfavorable contact stagger, this ground pulse is applied to the winding of the next relay and, unless minimum operating time definitely exceeds maximum bunching time, may P1

--..i.

------------

* (NI

111:i__

* THE CONTINUITY BREAK CONTACT IS NOT NECESSARY IF SEPARATE INPUT LEADS, Pi ANO Pz, ARE PROVIDED. Flg. 11-27 Recycling Sequence Circuit Employing One Relay per Stage

The Design of Switching

272 operate the relay falsely. of two input leads.

Again, this hazard can be avoided

Circuits

by the use

Consideration of the operating plan indicates that some method of pre-operating the first relay is necessary for the circuit to function. This can be provided by a series path to ground through break contacts on all relays except the first, as shown by the top row of contacts on the circuit of Fig. 11-27.

If the recycling requirement is not specified, the circuit can be modified by substituting transfer chains for the separate makes and breaks shown for the operating paths in the circuit and for the output paths. The design of sequence circuits to fulfill other sets of requirements may be carried out in a manner similar to the examples given above. Usually it is found that one of the conventional counting circuits may be used with little or no modification as a basis on which the required circuit arrangement may be constructed.

PROBLEMS 11-1

FOR CHAPTER

A sl.x - pulse relay counting circuit Is to be constructed, using pulse - frequency dividers of the type shown In Fig. 11--4 as a basis for the design. The counter Is to recycle following the termination of the sixth pulse. Relays to control the pulse dividers may be used where necessary. Assume a slow pulse rate. (This can be done with six relays Including two pulse dividers.)

11-2 Design a three-pulse relay counter, employing cult Is to recycle after the third pulse. 11-3

11

stages of "prime-pairs".

The cir(Sl.x relays)

Design a counting circuit to count a maximum of five pulses using the minimum numbers of relays. The circuit should not recycle after the !Uth pulse. (Four relays)

11--4 A fast-acting counting circuit to count a maximum of six pulses is to be designed. The circuit arrangement Is to include five relays but no resistors, since "'shuntdown" operation would make the circuit undesirably slow. The circuit should lock upon the final pulse and furnish an output Indication until released.

11-5

(a) On a main street In a certain city there Is a series of five traffic lights. These five lights are to be controlled by a relay circuit In such a manner that the green and red lights regulating the main street traffic are lighted for one and a half minutes and one minute, respectively. In order to Insure accurately-timed Intervals, a source of ground pulses of one second duration with a rate of two pulses per minute Is made available. The five lights are to be controlled progressively: that is, the lights do not all change at once; the light at the second corner should change 30 seconds after the light at the first, and so on. Design the required relay circuit. (This can be done with either depending upon the type of sequence circuit used.)

five or ten relays,

Circuits

for Counting

273

(b) A sixth traffic light ls added to the arrangement of part (a) at the corner followcontrol key ing the fifth light. This new light is normally off. When a non-locking is operated, this sixth light should function in a manner similar lo the other lights, following in sequence 30 seconds after the fifth light. However, the light should not interval. begin to (unction until the beginning of Its "green-light" A second operation

Add this relays.)

feature

of the key should discontinue to the circuit

operation

or the sixth tramc

of part (a). (This can be done with

light.

four additional

11-6

An interrupter make-contact operates continuously in a cycle of one second closed, one second open. Design a circuit lo deliver two complete one-second ground pulses start-key. The output pulses after momentary operation of a single make-contact should be the first two full pulses after the momentary start-key closure. Following the two output pulses, the circuit should automatically restore to normal. The startkey will not be reoperated during the acting time of the circuit. (This can be done with four relays.)

11-7

U a coin is tossed a number of times, It may be expected that the number of"heads" A relay circuit ls to be dewill be approximately equal to the number of "tails". signed to aid in the study of this phenomenon. To control this circuit, two non-locking keys (H) and (T) are provided, each with a single make-contact having one spring grounded. After each toss of the coin, one of these keys is operated: key (H) if the coin falls "heads", and key (T) U the coin falls "tails". U, after any toss, the total number of "heads" obtained since the beginning of the test run exceeds the number of "tails" by three, a lamp designated (3H) is to light. Similarly, if at any time the total number of "tails" exceeds the total number of "heads" by three a lamp designated (3T) is to light. The total number of tosses is not of interest to the circuit. After either the cleared, and the the operator falls either lamp (~H) llghted until the

(3H) or (3T) lamp lights, a new test run must begin. The circuit ls lighted lamp extinguished, by operation of a reset key. U, however, to clear the circuit and, Instead, operates key (H) or key (T) after or (3T) is lighted, an alarm lamp (ALM) ls lighted and remains reset key is operated. (This can be done with nine relays.)

SPECIAL PROBLEM This problem is useful and Instructive as a combined circuit design and laboratory construction exercise. Although the circuit is not specUically of the counting type, the principles covered in this chapter are useful in its solution. The problem ls to design and build the circuit portion of an electric combination door or safe lock. The complete lock consists of a relay circuit, a set of control keys, and two lamps, one representing an electromagnetic door latch and the other, an alarm bell. The key and lamp arrangement to be used is shown on the figure below. The circuit should employ no more than five single-winding general-purpose relays. The "lock lamp" conditions~

(L)

and the "alarm

lamp"

(ALM)

should

light

under

the following

The Design of Switching

274

Circuits

Lock lamp (L): When control keys (A), (B), (C), (D) are depressed and released in the sequence D-B-C-B-A. (The assigned sequence may be varied, as A-B-C-B-A, C-A-D-A-D, or any other sequence of five.) B-D-B-A-C, Alarm

lamp (ALM): l. U an Incorrect sequence of control keys is pressed. 2. U two or more keys are depressed simultaneously. 3. U the same key is depressed twice consecutively, unless specified sequence.

required

by the

When either lamp (L) or lamp (ALM) has lighted, no further operation of keys (A), (B), (C), or (D) should affect the output conditions. Operation of a release key (RL) at any lime In the cycle should restore the circuit completely to normal. Key closures of duration less than the acting time of a relay need not be considered as in1>Ut signals. L

A

,...

,...

-·

I

I

2

3

ALM

RL

D

C

B

I

4

5

6

7

8

9

10

II

L-

______

12

f1

BA;;ERY s_u_'.'~LY

Chapter

12

CODES AND TRANSLATING

cmCUITS

The input signals to large switching systems often represent either data which are to be processed by the system, or instructions which specify the procedures to be followed. These signals should be thought of as "information" rather than as remote-control means for setting switching devices. The information is recorded in the system by operating relays or other switching devices in combinations, and is passed from point to point by combinations of signals on groups of interconnecting leads. A set of combinations of elements in which each combination represents an item of information is a "code". Depending upon circumstances, one set of combinations may be more appropriate than another to represent particular information, and thus several codes for the same information may exist. A circuit arrangement which converts from one code to another is a "translating" circuit. This chapter will discuss various methods of representing information in the form of a code, and will also discuss the design of translating circuits. Numbers are used as a universal "language" for recording and transmitting information in switching systems. They are used in three ways: (1), to represent quantity; (2), to specify order; (3), as arbitrary symbols or designations. Numbers representing quantities are used extensively in problems solved by digital computers. There are also many instances where switching systems count objects or events and retain a record of the quantity. An example of numbers used to specify order is where a group of leads or switch terminals are numbered in the order of their physical location. The number of a particular lead or terminal then defines its location with respect to the other leads of the group. Numbers may be used as convenient arbitrary symbols to represent instructions given to a switching machine, or to identify entities which have no numerical characteristics of quantity or order. For example, telephone subscribers are assigned numbers. In some systems the numbers correspond to the location of this subscriber line on the terminals of selecting switches, and the number consequently contains intrinsic information which describes the location of the line. In other systems, however, there is no systematic correspondence between subscriber number and line location. The number is purely an arbitrary symbol identifying a particular subscriber. These telephone systems are provided with electrical "cross-reference" circuits in which are

- 275 -

276 recorded particular

The Design of Switching

Circuits

the line locations of all subscribers. This information for a line is made available when that line number is received.

Where numbers are used as an orderly set of designations, natural groupings of the designated items often correspond to a system of enumeration. Take for example a group of ten switches, designated 0 to 9, where each switch has ten terminals which in turn are designated 0 to 9. An individual terminal can be identified by a two-digit decimal number, where the first digit gives the number of a switch and the second digit gives the terminal on this switch. If there were ten groups of switches each containing ten switches, the system could be extended to three digits where the additional digit in the "thousands" place is the group designation. In practical examples, the number of terminals on a switch and the method of grouping often does not conform to a decimal system of designations. The system of enumeration is adapted to the size and grouping of the switches and may be not only non-decimal, but also mixed (that is, the "base" for each digit position is not the same). An example is a case where 10,000 terminals on the banks of twenty 500-point switches must be identified individually. The twenty switches are divided into five groups, each group containing 2,000 terminals and consisting of four switches, each switch containing 500 terminals. The switches are constructed so that the 500 terminals of each switch are divided into five groups of 100 terminals. These in turn are divided into ten subgroups of ten terminals each. A particular terminal is identified by specifying successively smaller groups and subgroups in which it is located. The first and largest group is one of the five groups of four switches each; the second is the particular one of these four switches. Having specified a particular switch containing 500 terminals, the particular terminal can be identified as being within one of the five groups of 100 terminals on this switch. Within this group, it is located in one of ten subgroups and is a particular one of the ten terminals. Thus the terminals may be identified by five -digit numbers, where the bases of the digits are 5, 4, 5, 10, and 10. A non-decimal numbering system of sp~cial importance in the switching art is the binary or base-2 system. As discussed in the previous chapter, in this system each "place" has but two values, 0 or 1. The importance of the system is that it may be represented directly by the two-valued elements used in switching. Thus, for example, a relay may be assigned to represent each place in binary notation and the relay may be operated or not, depending on whether the "value" of this place for the particular number is 1 or 0.

Codes and Translating 12.1

Circuits

HOW INFORMATION

277

IS REPRESENTED

The digit in each "place" of a multidigit number may be represented in switching circuits by providing apparatus which is capable of assuming a number of distinctive "positions" or "conditions". A single decimal digit may be represented, for example, by a ten-point switch which is set on the point corresponding to the digit value. When relays are used to represent multivalued digits, several relays are provided for each digit place and these relays stand operated or released in specified combinations to represent the digit values. For example, four relays may operate in sixteen different combinations (including all released), and ten of these combinations may be chosen to represent the which have defined meanings digits O to 9. Such a set of combinations is a "code". The "elements" of a code are those items which are available for choice in forming the combinations composing a code. In the above example, where four relays are operated in combinations to represent the decimal digits, the code is a four-element code, each relay corresponding to an element. Most of the discussion in this chapter is concerned with two-valued elements where one of the two "values" is specified as an active condition. Some of the two-valued elements used in switching for representing codes are shown in Table 12-1. Code Element

Active

Condition

Inactive- Condition

Relay

operated

released

Circuit input or output lead

grounded

open

Circuit input or output lead

battery lead

open

Gas tube

fired

extinguished

hole punched

no hole

Position

on paper

Table 12-1

Typical

on

Code Elements

In addition to numerical information, codes are also used to represent other forms of symbolism. For example, a five-element code (maximum 32 combinations) is used to represent the letters of the alphabet in teletype and other applications. Also, in a switching circuit where a variety of input control signal combinations each cause a different circuit response, a tabulation of the input signal combinations with the corresponding circuit actions is the equivalentof a code, and certain parts of the circuit design may be similar to that of translating circuits. Since codes representing the integers 0-9 occur most frequently in switching systems, the discussion in this chapter will be confined to these codes.

278 12.2

The Design of Switching SIMPLE

COMBINATIONAL

Circuits

CODES

Any set of ten different combinations of a number of elements assigned as a code to represent the integers 0 to 9. may be arbitrarily The combinations forming a code sometimes arise naturally due to particular circuit actions. Designation of For example, in the single decade counter of Relays Operated Digit Fig. 11-19, the digits 1, 2, ... 9, 0 are repre(l) l sented by the operation of either one or two of the six relays in the circuit. The combinations (2) 2 form a code for representing the ten digits as (3) 3 shown in Table 12-2. (4) 4 At least four code elements are neces(5) 5 sary to represent the ten 'integers Oto 9. Four (5), (6) 6 two-valued elements will give a total of sixteen combinations, and any ten of these may be used (1), (6) 7 to form the code; but it is desirable to assign (2), (6) 8 the combinations to represent the integers in a 9 (3), (6) systematic manner. Most of the standard codes in switching systems are "additive", that is, the (4), (6) 0 elements of the code are given numerical designations and the combinations are chosen so Table 12-2 Code for the that the sum of the designations of the activated Single-Decade Countlng Circuit or Chapter 11 elements (for example, operated relays) gives the value of the decimal digit represented by that combination. This makes the code combinations easy to remember, a characteristic of considerable practical importance. Four-element codes of this type are shown in Table 12-3, in which the designations of the code elements are given at the top of each code column and the elements activated for each digit value are tabulated in the column. When applying these codes in practical problems, special consideration must often be given to the combination to be used for zero. In many cases it is satisfactory to let the "blank" condition of all relays released represent zero. However, the zero condition is then the same as the normal condition when the circuit is idle. When it is desired to let zero be represented by an active combination, the additive combination for ten may be used or one of the unused combinations may be assigned arbitrarily (e.g., 1-4 in the 1-2-4-5 code). The choice of a combination to represent zero often depends on whether zero occurs as the first or the last of the integers 0 to 9. For example, if digits are represented by sequences of pulses, the digit "0" will be represented by ten pulses. If it is then to be converted to a code it probably should be represented by an active combination. Each of the four-element codes has characteristics which makes it more suitable for some applications than others. The 1-2-4-8 code

Codes and Translating

279

Circuits

Code Elements Digit

1-2-.C-8

1-2-.C-5

1-2-.c-6

1

1

1

2 3

2

1 2

2

1-2

1-2

1-2

4

4

4

4

5

1-4

1-4

6

2-4

5 1-5

7

1-2-4

1-6

8

8

2-5 1-2-5

2-6

9

1-8

4-5

1-2-6

o•

2-8

1-4

4-6

"Digit O moy be represented in any of these codes.

Table

6

by a blank combination

12-3 Four-Element

Addltlve Codes

corresponds to the binary system of notation and is often called the "binary" code. The values of the elements are powers of 2, that is. 2°, 2 1 , 22, and 23; and these values can be added together in only one way to give a particular digit. Thus there is no ambiguity in determining additive combinations. Although only ten combinations are used, the code retains its additive nature for all of the sixteen possible combinations.• It can easily indicate the odd or even character of a registered number since the element "l" appears only in odd digits. When "0" is considered as the first one of the ten integers, the 1-2-4-5 code has the characteristic that the digits below five (0, 1, 2, 3, and 4) do not contain code element "5" in their combinations, while all of the remaining digits (5. 6, 7, 8, and 9) contain element "5". Note that the non-additive combination of 1 and 4 for zero (instead of 1 + 4 + 5 = 10) conforms with this use. When "0" is considered as the last of the integers, the 1-2-4-6 code has similar characteristics. The element "6" can be used to distinguish between the first five integers (1, 2, 3, 4, and 5) and the last five (6, 7, 8, 9, and 0). teristic

The four-element codes of Table 12-3 have the common characthat a single "error" may convert one valid combination of the

• Thls supposes that the blank condltlon table, 2-8 ls often used.

ls used for zero although, as lndlcated ln the

280

The Design of Switching Circuits

code to another valid combination in the code. In the case of relays, an "error" consists of either the failure of a relay to operate when it was intended to operate, or the false operation of a relay when it was intended to remain unoperated. For example, in the 1-2 -4-5 code, if the intention is to indicate digit 3 by operating relays designated ( 1) and (2), an open lead which causes relay (1) to fail will falsely indicate digit 2. Again, if digit 3 is intended and a trouble ground causes relay (5) to operate as well as relays ( 1) and (2), then digit 8 is falsely indicated. 12.3 SELF-CHECKING CODES

Two-out-of-Five

Biquinary

Digit 00-5-0-1-2-3- .. 0-1-2-4-7 By choosing the combinations of a code so that a single 00-1 0-1 error produces a combination 00-2 0-2 2 which is not valid in the code, the occurrence of an error can 00-3 1-2 3 be detected. Codes of this nature 00-4 4 0-4 are called "self -checking" codes. 5-0 1-4 5 At least five elements are necessary to provide ten self-check5-1 2-4 6 ing combinations for the decimal 5-2 7 0-7 digits. A five-element self5-3 1-7 8 checking code is shown in Table 5-4 12-4. This uses the ten combina2-7 9 tions of five elements taken two 00-0 4-7 0 at a time, and is known as a twoout-of-five code. The elements Table 12 --4 Self-Che eking Codes are designated 0, 1, 2, 4, and 7, and the combinations are additive except for zero which is 4 and 7. With this code, any single error resulting in the false operation or failure of a relay to operate will leave either one or three relays operated and can be easily detected.

A second self -checking code shown in Table 12-4 uses five relays designated O to 4, and two additional relays designated 00 and 5, which operate respectively when the digit is below six or above five. This corresponds to the biquinary system of notation and is useful in the design of circuits to perform arithmetic calculations. This code is, therefore, often used in computers. The two codes of Table 12-4 are sometimes abbreviated by omitting the "zero" elements. This, of course, destroys the self-checking nature of these codes but reduces the number of code elements required. This simplified representation of the biquinary code is widely

Codes and Translating

Circuits

281

used in computers since it produces quite simple circuits for performing addition. For this reason it is preferred over either the fourelement codes which require one less code element, or the two-out-offive code which is self -checking but not suited to calculating circuits. The full seven-element biquinary code is both self-checking and convenient for use in calculating circuits. 12.4 THE THEORY OF ERROR-DETECTING AND ERROR-CORRECTING

CODES

The self-checking quality of the two-out-of-five code depends on the relations between the various combinations chosen to represent the items of the code. The combinations are so related that at least two errors (false operation or non-operation of a re-lay) are necessary to convert one valid combination to another valid one in the code. The principle can be generalized and extended to detect not only the occurrence of single errors but also, by increasing the number of elements (relays or leads), to detect multiple errors and to indicate which elements have failed when an error occurs. That is, a code can be made not only "self-checking" but also "self-correcting," and the errordetecting and error-correcting ability can be extended to any number of simultaneous errors. This, of course, is accomplished at the expense of increasing the number of elements in the code. In order to explain the general principle, let the "distance" between two combinations be defined as the number of changes of relay positions (operated and released) necessary to convert one combination to another. For discussion purposes, combinations may be represented as in switching algebra, where 0 and 1 represent' respectively:the operated and released positions of a relay, and the symbols are written in a prescribed order for a given set of relays. For example, the particular combination of three relays, A, B, and C, in which A is operated, B released, and C released may be written 011. To illustrate the meaning of "distance" the combinations 111, 001, and 010 are all at a distance of one "change" from the combination 011. The distance between combinations 011 and 101 is two changes, and so on. • A single error in a given combination will produce a combination at a distance of one from the original. Two simultaneous errors will make this distance two; three errors will produce a distance of three; and so forth. In order to produce a set of combinations which is self checking for single errors, it is necessary only to select a set of valid combinations in which the distance between combinations is at least two. An error then produces a combination which can be recognized as not being one of those used in the code. This is illustrated for three relays in Fig. 12-1 which represents the eight possible combinations in

The Design of Switching

282

Circuits

their proper relation by the corners of a cube.• Distances between combinations are measured along the edges of the cube. The four combinations, 000, 011, 101, and 110, indicated by heavy dots in the figure, are all at a distance of two from each other. To produce a "single-errorcorrecting" code, the combinations are chosen so the minimum distance between combinations is three. A single error then produces a combination at a distance of one from the intended combination but at a distance of two or more from any other combination of the code. Thus single errors can be corrected by changing to the nearest code combination. This is illustrated in the cube of Fig. 12 -1 by the pair of combinations 000 and Fig. 12-1 111, which are separated by a disA Cube Representing the Eight tance of three. A single error in 111 Combinations of Three Elements results in one of the combinations 110, 101, or 011, each of which is closer to 111 than to 000. Only a single pair of self-correcting code combinations is possible with three elements, but these may be any of the pairs at opposite ends of a diagonal of the cube. To represent a larger number of items by a selfcorrecting code, the number of elements must be increased.

--------

110

Ill

A set of single-error-correcting combinations fails in the presence of two errors. The occurrence of two errors produces a combination which is nearer to some combination other than the correct one, and thus an attempt to correct in this case results in the combination being "corrected" improperly. If the minimum distance between code combinations is increased from three to four, a double error will bring the combination to a point halfway between two (or more) of the code combinations and thus the double error can be recognized. Codes having a minimum distance of four are able to correct single errors and todetect the presence of double errors. By increasing the distance between combinations, codes can be developed to correct or detect the presence of any number of errors. The "distance" concept is helpful in understanding the theory of error-detecting and error-correcting codes. However, in practice the • Similar figures for a larger number of elements correspond to cubes of higher dimensions. The four-dimensional "hypercube," for example, has slxteen corners correswere ponding to the combinations of four elements. These geometrical representations discussed in Section 8.7, the appendix to Chapter 8.

Codes and Translating

283

Circuits

distance requirements are arrived at indirectly, and it is difficult to see intuitively that the various combinations have the desireddistance relations. Mathematical proofs are not given here but actual trial of a few examples will show that the codes do permit errors to be detected and corrected, and therefore the distance requirements are met.

X A B C D 1

1

2

0

3

0 1 1 0 1

1 0 1 1 0 0

4

0 1 In developing a code, a set of combina6 1 0 1 0 tions having the required characteristics can 7 1 0 0 1 be chosen so that a system of "even-or-odd" checks of the elements indicates the presence 8 0 1 0 0 0 of errors. Certain elements of the code may 9 0 0 1 1 1 carry the "information" while other elements 0 1 1 0 provide the checks and are chosen according to the combination of the information ele11 1 0 0 1 ments. This is illustrated by Table 12-5, 12 0 0 1 0 0 which shows a check element X and four in13 1 0 0 1 formation elements A, B, C, and D. All of the sixteen possible combinations of A, B, C, 14 0 0 0 0 and D appear in the table. For each combina15 0 0 0 0 1 tion, X is chosen to be O or 1 to make the 16 1 0 0 0 0 total number of 0' s an even number. A single error in any element will result in an odd Table 12-5 number of 0' s (either 1, 3, or ~) and thus can Self -Checking be easily detected. Ten of the combinations Combinations have exactly two 0' s, and these are used in practice to represent the decimal digits in the two-out-of-five self-checking code which has been discussed. This permits a somewhat better checkthan a straightodd-even check since two errors must be of an opDigit 0 1 2 4 7 posite nature to escape detection. That is, one relay must falsely release and another 8 1 0 1 0 falsely operate. In the additive 0-1-2-4-7 2 0 1 0 1 code in Table 12-4, the zero element is 5 0 0 purely a check element since its inclusion in an additive combination does not affect 0 1 1 0 the total. 2 0 1 0

5

0

" 3

1 0 0

1

Check

0 0 0

0

Table 12-6 Scheme

A Correcting

To indicate the location of an error so that a coding system ls self - correcting, additional odd-even checks are necessary. One method is to check a group of coded digits by providing a supplementary

284

The Design of Switching

Circuits

"check digit'' whose combination of O's and l's is determined from the group of digits which it checks. The method is explained by referring to Table 12-6, which shows the development of a check digit for the sixdigit number 825423. The six digits are represented in the 0-1-2-4-7 code as shown in the table.The O's and l's of the check digit are chosen for these particular six digits so that there is an even number of O's in each of the five columns of the table. Errors are located by a combination of "row" and "column" checks. Each digit corresponds to a row in the table, and a failure of the "even" (or two-out-of-five) check of zeros in this row indicates an error in one of the elements of this row. Failure of the "even" check of the seven elements of a column indicates an error in that column. Thus the element at the intersection of the row and column is in error and if O, should be cha11ged to 1, or if 1, should be changed to O. This illustrates the basic principle of self-correction. Each element is checked by several checks of the odd-even type in such a way that no two elements are checked by the same combination of checks. In the above example, each row and column of the table has its own oddeven type check, and the intersection of a row and column indicates a particular element. In the correcting scheme illustrated by Table 12-6, a single check digit can serve for a series of digits of any length. The possibility of a double error, however, increases as the number of digits increases. The scheme recognizes the presence of a double error but cannot correct or in some cases even determine which digits are in error. Two errors in general cause two row checks and two column checks to fail and cannot be corrected with certainty. If two errors occur in the same column, two row checks and no column checks will fail. Thus there is no indication of which column is in error. Similar effects occur for double errors in the same row. • A coding method, developed by R. W. Hamming, will now be described which permits individual code combinations to be corrected for single errors. By this method, seven code elements can be arranged in sixteen self-correcting combinations. Four of the seven elements are information-carrying elements and may be used in any of their sixteen possible combinations. The remaining elements are check elements and their value (0 or 1) is determined according to a system of odd-even checks. The system is explained by Table 12-7, where the seven positions (columns) correspond to the elements of the code. Positions 1, 2, and 4 are occupied by check elements, and the remaining positions, 3, 5, 6, and 7, contain elements designated A, B, C, and D which carry the information. Three odd-even checks are applied to subgroups of the seven elements. As indicated by asterisks in the table, the first check, which

Codes and Translating

Circuits

285

Code Positions Checked 2

3

4

Check

X

X

A

X

Xl

*

* * *

'

X2 X4 Table 12-7

5 B

6 C

7

Position

D

Code

* * * * * * * *

Check Arrangement

for Sell-Correcting

Code

will be referred to as the Xl check, is applied to the set of four elements located in positions 1, 3. 5, and 7. For a particular code combination of the A, B, D elements, the value of the check element in position 1 is chosen to be O or 1 so that the total number of O's in the four positions is even. The second check, X2, applies to positions 2, 3, 6, and 7. Element 2 is made O or 1 to satisfy an even check of O's in these positions. In a similar manner the third check, X4, is made on positions 4, 5, 6 and 7. Before explaining the significance of into check groups, the application of the trated using the combination 0101 in the values of the elements in the three check follows: Position:

this subdivision of elements above process will be illusinformation positions. The positions -are determined as

1 2 3 4 5 6 7 BCD

A

•

0

• • •

-

1 0 1

• • • • ••••

To have an even number Position 1 must be Position 2 must be Position 4 must be

of O's: 0 1 0

Since A is O, and B and D are 1' s, the first position must be filled with an O in order to make an even number (two) of O's in the four positions 1, 3, 5, and 7. In a similar manner positions 2 and 4 are found to be 1 and O respectively. The complete seven-element combination, when the three checks are included, is 0100101. The significance of this subgrouping of elements into check groups is that each of the seven elements is in a different combination of check groups. For example, element 1 is in check group Xl only, element 2 in X2 only, element 3 in Xl and X2 but not in X4, and so on, element 7 being the only element in all three check groups. To determine. whether an error has been made in a particular seven-element code combination, the three odd-eve!\ checks Xl, X2, and

286

The Design of Switching

Circuits

X4 are made. If all checks are even, then no error has been made. If one or more checks indicate odd, then the error was made in the position checked by this combination of checks. The checks which fail indicate in an additive code the position of the element which is in error. For example, a failure of Xl and X2 indicates that the position 1 + 2 = 3 is in error; failure of X2 and X4 indicates position 6; all three, position 7; and so on. To show the correction procedure, assume that the combination 0100101 which was developed above is received with an error which results in the combination 0100001. The checks are applied as follows: Position:

1 2 3 4 5 6 7

Number of O's in check:

0 1 0 0 0 0 1

•

• • •

• •• •• •• •

Xl = Odd X2 = Even X4 = Odd

Checks Xl and X4 show an odd number of 0' s and therefore the element in the fifth place is in error. Since this is 0, it should be changed to 1 to give the correct combination. Note that this procedure will correct errors in the check elements in positions 1, 2, and 4 as well as the information elements. If only one of the checks gives an odd count the error is in the corresponding check position. A table of the enXXAXBCD tire sixteen self-correcting combinations based on the above procedure 1 1 1 1 is given in Table 12-8. 1 0 0 0 0 2

3 4 5 6 7 8 9 10 11 12 13 14 15 16

1 0 0 0 1 1 0 1 0011001 1 1 0 0 0 1 1 0 0 1 0 0 0 1000011 0 0 1101001 0 0 0

0 1 0 1 1 0 0 0 0 1 0 1 0 0

0 0 1 0 1 0 1 0 1 0

0 1 0 1 0

1 0

1 0

0

0

0

0

Table 12-8 Seven-Element SlngleError-Correctlng Combinations

Ten of the sixteen combinations in Table 12- 8 may be arbitrarily chosen to represent the integers 0 to 9; if desired, the information elements A, B, C, and D may be assigned values according to one of the four element additive codes. The code may be varied by substituting odd checks for even checks in any one or more of the three checks. If more than sixteen self-correcting combinations are required, the number of information elements of the code must be increased and this in turn necessitates an increase in the number of check elements. The addition of a fourth check symbol in position 8, following position 7, will

Codes and Translating

Circuits

287

permit the code to be extended to elements in a total of fifteen positions, eleven of which are information positions whose elements may assume any desired combination. This fourth check, X8. will be associated with positions 8 through 15 as a group, and the previous checks, Xl, X2, and X4, will be extended through positions 9 to 15, repeating the same pattern as in positions 1 to 7. In this manner the code may be extended to as many positions as desired, n-check elements being able to indicate single errors in (2n - 1) positions. The correcting feature of this code fails when two errors occur. Applying the check procedures will result in one or more odd checks, but applying the indicated correction will result in a combination which is not the one originally intended, although it is one of the established combinations of the code. This can be verified by trial, using any of the combinations in Table 12-8. One way to indicate the occurrence of two errors is to add an eighth element to the seven-element code. This "master check" element, M. may be placed in position "zero" ahead of position 1. It is made 0 or 1 in a particular combination so that the total number of O's in the eight positions is even. Failure of this check then indicates the occurrence of a single error. The correcting procedure may be deduced as follows: If Master Check of All Elements is:

And If Group Checks Xl, X2, X4 are:

Even

All Even

No Errors

Odd

Some Odd

Single Error: Correctable

Even

Some Odd

Double Error: Not Correctable

Odd

All Even

Error in Master Check Position: Correctable

This eight-element code gives erroneous three or more errors occur. Further reliability, require the use of additional check elements.

Then the Indicated Condition is:

indications only when if demanded, would

Error-correcting codes of the above type permit the correct information to be reconstructed on a digit-by-digit basis when errors occur. For example, when information is recorded as punched holes in a paper tape, errors due to mutilation or flaws in the paper could be corrected at the time the information is read. Various advantages may be illustrated by a circuit arrangement indicated in Fig. 12-2. The input consists of eight leads which include three check leads, four information leads, and a master check lead for double-error detection.

The Design or Switching Circuits

288

These leads are normally grounded in the combinations or a code which corrects single errors and detects double errors. When no errors o~cur, the input combinations are repeated on the output leads. In case a single error occurs in a combination presented to the input leads, the correct combination will be produced on the eight output leads and, in addition, some combination or the Xl, X2, and X4 leads is grounded to indicate which input lead is in trouble. Thus the apparatus on the output could continue to !unction properly and, i! desired, the trouble information could be recorded and the trouble cleared at some later time. In case or two errors in the input combination, the double-error lead (DE) is grounded and no other output leads grounded. This indicates an inoperative condition which cannot be automatically corrected. The output leads, instead or merely repeating the input code, could That is, present the information in any other code arrangementdesired. the circuit may translate as well as correct. Theoretically, this circuit requires only eight relays, one on each input lead, and may be designed by ordinary combinational methods. However, it is evident that the contact networks will be rather elaborate, and auxiliary relays are necessary if relays or practical size are to be used. 2

3 4

INPUT COMBINATIONS

5 6 7

2 3

ERROR

4

CORRECTING ANO TROUBLE INDICATING CIRCUIT

5 6

CORRECTED OUTPUT COMBINATIONS

7

8

8

w a

~

SINGLE ERROR IDENTIFICATION

Flg. 12-2 Error-Correcting

12.5

'---t DOUBLE

ERROR

INDICATION

and Trouble-Indlcatlng

Circuit

TRANSLATING CIRCUITS

A translating circuit is one which converts or translates information received in one code to the same information expressed in some other code. As indicated in the block diagram of Fig. 12-3, the circuits are basically combinational. Input conditions operate relays according to one code, while contacts on these relays ground (or otherwise activate) output leads according to the combinations of a second code. A given input combination always produces the same output combination

Codes and Translating

Circuits

INPUT { CODE LEADS

Fig. 12-3

289

TRANSLATING CIRCUIT

General Block Diagram

}

OUTPUT CODE LEADS

of a Translating

Circuit

regardless of the sequence in which input combinations occur. Translating circuits may work between any two codes, the circuit usually being a "one-way" device. That is, a circuit which translates from a code Kl to a second code K2 will not also perform the inverse translation of K2 to Kl. In some translating circuits the input leads"directly control relays whose only purpose is to perform a translating function by means of a network of contacts. In other cases the relays carrying the translating network may be intimately involved in some other function, and render it difficult to isolate actual input leads to the translating circuit. The input conditions then are considered to be the operation of the relays rather than conditions appearing on specific input leads. An example might be a counting circuit, with the sequence shown in Table 12-2, arranged to ground one of ten output leads according to the number of pulses counted. The counting function is the means for causing the relays to establish successive combinations for counting the input pulses, while the translating circuit network causes these combinations to ground one of ten output leads. 12.6

DESIGN

TECHNIQUES

FOR TRANSLATING

NETWORKS

Although the design of a translating network involves only the usual principles of contact network design, certain characteristics of the code combinations and the practical requirements which must be met by the network indicate lines along which circuits may be developed. If the input code combinations include all of the possible c01nbinations in which the relays may operate, then the action of the circuit for all conditions is completely specified and the design of a network is entirely routine. However, when some of the possible combinations are not used, as is more often the case, decisions must be made as to whether these invalid combinations may represent closed paths in the contact network. By taking advantage of the invalid combinations it is often possible to reduce the required number of contacts in the network. Consider as an example any one of the four-element codes of Table 12-3 where "0" is represented by operated relays. Of the sixteen different combinations in which four relays may operate, one is the "normal'' or "all relays released" condition, ten combinations represent the decimal

290

The Design of Switching

Circuits

digits in a four-element code, and five combinations remain unused. In most circuit applications, these five invalid combinations never occur in normal circuit operation, so that allowing the network to close one or more of its output leads for these combinations is theoretically satisfactory and may permit a simpler network to be devised. An invalid combination will occur only because of a trouble condition such as a defective relay or contact, or trouble crosses, grounds, or opens in interconnecting leads. ll the nature of a circuit and its surroundings is such that false combinations are extremely unlikely, or if the occurrence of a false combination causes only local reactions, the best procedure is to design the simplest possible network which satisfies the code conditions and to permit output leads to be open or closed as convenient for the invalid combinations. Whe.re false output conditions will cause objectionable reactions in associated circuit units, or will cause appreciable portions of a system to be made inoperative for a period of time, then it is advisable to impose the strict requirement that output leads be grounded only when the input combinations are those specified in the code. Although this may result in more complex contact networks than when the less strict requirements are imposed, the design work is often easier since the requirements are completely stated and it is not necessary to try various arrangements with invalid combinations to determine which produces the simplest network. The extension of the requirements to include a signal on an alarm output lead when a false combination occurs is obvious. The self-checking codes, of course, make these trouble protection measures more effective. Another requirement which may be necessary and which influences the choice of contact network arrangements is that ungrounded output leads may not be crossed during normal circuit actions. This condition is imposed when crosses or "back-ups" between ungrounded TRANSLATING CIRCUIT

OTHER CIRCUIT

r--------1

--!--

INPUT

Fig. 12-4

An Effect of Crossed Output Leads

Codes and Translating

Circuits

291

CHECKING NETWORK

TRANSLATING NETWORK

l

OUTPUTS

L--,---J I

RELAY COILS

INPUTS Fig. 12-5

Use or a Checking Network in Translation

leads may cause false actions in circuits connected to the outputs. An example is indicated in Fig. 12-4 where a particular input condition to the translating circuit grounds output lead X as shown and also connects leads Y and Z together. If relay (Y), for some reason, has been previously operated and locked, as indicated, the locking ground will "backup" through lead Y to lead Z and falsely operate relay (Z). This trouble can be corrected by redesigning the translating network to eliminate crossing of leads. An alternative solution which does not involve the translating circuit is to remove the grounding condition in the external circuit. In the example shown, this can be done by using a separate locking winding on the relay, or by operating and locking the relay through a make-before-break transfer.

If it is necessary that output leads should not be crossed, care must be exercised in combining the networks of individual output leads. Satisfactory circuits are determined by inspection and analysis, assuming that all possible situations occur, and checking for back-up paths. One method of preventing the appearance of grounds on output leads in case of a false input combination is to provide a ''checking" network as indicated in Fig. 12-5. The checking network is closed for all valid input combinations of the code, but is open for all other combinations. The translating network can then be any network which closes the output leads according to the required code. Sometimes the form of the checking network is obvious, as in the case of the two-out-of-five code in Table 12-4 where a simple symmetric network may be used. The function of the checking network is to close the circuit path only when the combination in which the relays stand is one of those specified in the code. Whether this is, in reality, a "check" of the validity of the information depends on the nature of the code. To illustrate: when a ''self-checking" code is used, a single error produces a combination

The Design of Switching

292

Circuits

which is not in the code and consequently a checking network will detect the error by opening the circuit path. However, for other codes, where certain errors convert one combination to another in the code, an error is detected only when the resulting combination is one of those not used in the code. A checking network of this type can only determine that the registered combination is one of those specified in the code and not that it is the one originally intended. OUTPUTS 8

2

9

0

3

4

6

5

INPUTS Fig. 12-6 Translating

12.7

Circuit

EXAMPLES ILLUSTRATING

!or Pulse Counter

Relays

DESIGN TECHNIQUES

Several of the forms a translating circuit may take will be illustrated and compared by developing circuitarrangements which perform the same translation. The codes are based on the operating combinations of the counting circuit given in Table 12-2. The translating circuit consists of a network of contacts on the six relays which grounds one of ten output leads according to the digit registered on the relays. A simple circuit can be easily developed from an analysis of the conditions shown in the table. The output lead for digit 1 is grounded when relay (1) is operated, provided also that relay (6) is released. The lead for digit 7 is grounded when (1) and (6) are both operated. Therefore, a path through a make on (1) and a break on (6) is closed for digit 1, while a make on (1) in series with a make on (6) is closed for digit 7. Similar reasoning gives paths for the other eight digits, each path containing a make or a break on (6). The ten paths are combined to use a single transfer on (6) in the contact network shown in Fig. 12-6.

If a trouble condition causes the relays to operate improperly, the contact network of Fig. 12-6 does not protect against false output signals. For example if relays (1) and (2) are operated at the same time, output leads 1 and 2 are both grounded. Whether this circuit is satisfactory will depend on the likelihood of false action of the relays and the consequences of false signals given to the connecting circuits.

Codes and Translating (I)

Circuits

293

(2)

(3)

(4)

Fig. 12-7 Series Chains for Translating

(5)

(6)

Digits 1, 3, and 7

Although this is not a self-checking code, some degree of checking is possible. One procedure is to construct for each output lead a path which passes through a series chain of six contacts, one on each relay - a make or break depending on whether the relay is operated or released. Each path is closed for only the correct one of the 64 possible combinations in which the six relays might operate. Typical paths are shown in Fig. 12-7 for digits 1, 3, and 7. A complete circuit for all digits is shown in Fig. 12-8 where individual pathshave been combined. This circuit opens all output leads in the event a trouble condition causes the relays to operate in some combination not in the code. A third type of network for this example may be developed according to the plan of Fig. 12-5 where the checking network closes only for valid combinations of the code. Analysis of the code in Table 12-2 shows that the first five relays are operated one at a time. An indication that exactly one of these relays is operated is sufficient to show that the relays are operated in one of the correct combinations of the code. The one-out-of-five symmetric circuit shown in Fig. 12-9 will give this check and may be placed in the ground lead of the translator, Fig. 12-6. The resulting network grounds an output lead only if the six relays are (I)

(2)

(3)

Fig. 12-8 A Translator

(4)

which Prevents

(5)

(6)

False Outputs

The Design

294 (I)

(2)

(5)

(4)

(3)

of Switching

-~

Circuits

(6)

D _JD: ~---e~ - - - -

.J

Fig. 12-9 A Check Circuit for Use with Fig. 12-6

operated in one of the ten correct combinations of the code. Note, however, that this circuit checks only single errors in the first five relays and does not insure that the particular grounded output lead is actually the one intended. For example, if digit 7 is intended and relay (6) fails to operate, the combination for digit 1 is established and lead 1 is • grounded. The circuit of the combined Figs. 12-6 and 12 -9 does not effect a great saving in contactsprings over Fig. 12-8, but distributes the spring load more uniformly over the relays. A possible advantage of Fig. 12-8 is that crosses between output leads cannot occur since all paths are disjunctive. The other arrangement, however, will cross two or more output leads under certain trouble conditions. For example, if relays (1) and (2) are erroneously operated at the same time, no output leads are grounded since the symmetric circuit, Fig. 12-9, is open; but two pairs of output leads are crossed, lead 1 to lead 2 and lead 7 to lead 8 (see Fig. 12-6). Choice between the various circuit arrangements for this translating problem will depend on the particular conditions under which it is used. circuits

The design of translating circuits can be further which perform the inverse translation to that OUTPUTS

illustrated by above. These OUTPUTS

OUTPUTS 6

1

MULT. TO RELAYS 6, 8, 9 ANO 0,

~

6

J

J

INPUTS

INPUTS

A

Fig. 12-10 Three Arrangements

B for Translating

C

Digits 1 and 7

Codes and Translating

Circuits

295

circuits have ten input leads which are grounded one at a time to cause six output leads to be grounded one or two at a time. Referring to Table 12-2, the "Digit" column will now correspond to inputs, and the "Relays Operated'' column to outputs. If ten relays are used, one connected to each input lead and designated accordingly, the relays (6), (7), (8), (9), and (O) must each ground lead 6 and one of the leads 1, 2, 3, 4, and 5. Relays (1) to (5) require one make-contact each, while relays (6) to (0) each require two. These are connected as shown in Fig. 12-l0A for the (1) and (7) relays. The fact that relay (1) and the similar relays (2) to (5) each contain only a single make suggests that these relays may be omitted and the corresponding input leads connected directly through to the outputs. This scheme, shown in Fig. 12-10B, has the possible disadvantage that operating relay (7) places ground back on input lead 1. This can be eliminated by a break-contact on the relay as in Fig. 12-lOC. A translating circuit using this scheme requires five relays. A circuit arranged according to Fig. 12-l0A can be made to discriminate against false input combinations by providing a one-outof-ten symmetric circuit on the ten relays and supplying all grounds through this network, as in Fig. 12-5. A similar check cannot be made with Figs. 12-10B and 12-lOC since relays do not appear on all inputs. 12.8

FURTHER

EXAMPLES

BASED ON STANDARD

CODES

The general procedure for the design of a translating contact network for combination codes will be illustrated by a circuitwhich translates from the 1-2-4-5 four-element code to the 0-1-2-4-7 self-checking code. Four relays designated 1, 2, 4, and 5 operate according to the

Digit

(1)

(2)

,..,

(51

LO

L1

1

0

1

1

1

X

X

2

1

0

1

1

X

3

0

0

1

1

"

1

1

0

1

5

1

1

1

0

6

0

1

1

0

7

1

0

1

0

8

0

0

1

0

9

1

1

0

0

Input Relays

0

0

1

0

Output Leads Grounded

L2

L..

L7

X X

X X

X

X

X X

X X X

X X

X

X

1

Table 12-9 Comblnatlons for Translatlng

X

X

Circuit

The Design of Switching Circuits

296

four-element code. Contacts on these relays are then required to ground five output leads (LO, Ll, L2, L4, and L7), two at a time in the proper code combinations. Since a large portion of the sixteen possible input combinations are used, the methods of switching algebra may be employed to advantage in developing the network. (Eleven combinations are used, ten for the digits O to 9 and an eleventh for the normal condition of all relays released.) In order to introduce a partial safeguard against output errors, the use of invalid combinations will not be allowed. The code combinations are given in Table 12-9, where the symbols O and 1 represent operated and released relays respectively. The two output leads grounded for each digit are indicated by X's in the table. The first step in the design is to develop a network for each output lead individually by considering the conditions which must ground this lead. For example, output lead LO is grounded for digits 1, 2, 4, and 7. The corresponding relay combinations are tabulated and the circuit for this lead developed by switching algebra as follows:

.!.!..i.§.

~!

f(LO)={~ 1 1 0 1!}=[2' 1 0 1 0

+ (1 + 4')(1' + 4) + 5'](1'

+ 2 + 4') •

To illustrate further: lead L4 is grounded for digits 4, 5, 6, and O; and the circuit is developed as follows:

{!! ~

f(L4) = 0 1 1 0~}= 2' + (4 + 5')(4' + 5) • 0 1 0 1 The circuits for the remaining leads, Ll, L2, and L7, are developed in a similar manner, and the five networks are then examined for possible simplifications. By properly arranging the order of the elements in the individual networks, and by making use of the disjunctive nab.lre of transfer contacts, the five networks may be combined as in Fig. 12-11. In this figure the paths of f(LO) and f(L4) developed above are shown in heavy lines for easy reference. It should be noted that translation from a simple combination code to a self-checking code, as in this example, does not permit checking to a greater extent than is possible in the original simple code. An error in the four-element input code which converts a code combination to another valid combination will simply produce a two-out-of-five output combination indicating the wrong digit. Errors which cause more or

Codes and Translating

Circuits

297

less than two output leads to be grounded correspond to invalid input combinations which can be detected without performing a translation. The purpose of translating to a checking code would be to detect troubles during later circuit actions. In some translations a characteristic of the code combinations may indicate a procedure for designing the network. For example, in translating from the two-out-of-five code to a code consisting ofground on one of ten output leads, the translating network may consist entirely of make-contacts, since only two of the five input relays are operated at any given time. A symmetric check network may be used ahead of the translating network to prevent false output grounds if a trouble condition causes more than two relays to operate. A circuit of this kind is shown in Fig. 12-12. The translating network may be arranged in a

LI

2

4'~ I'

LO

s' I

---

OUTPUT LEADS

4')-!

2•~

11 --- 4

s L2

S'

o 2

4•L

:.r

2'

4

-

0--.:,

0

I

0

5

L7

L4

Flg. 12-11 Network for 1-2-4-5 to 0-1-2-4-7 Code Translation

The Design of Switching

298 CHECK NETWORK

TRANSLATING NETWORK

·1:: ·1:: ·1:;

4

I

-½-

10---00

Fig. 12-12

Circuits

Two-out-of-Five

LI L3 LB L2 L6 L9

OUTPUT LEADS

L4 L5 LO L7

Code Translation

number of ways which result in different distributions of spring loads on the five relays. The one shown in the figure distributes the overall load of the check and translating networks by favoring in the translating network those relays which have heavy loads in the symmetric check network. The check does not guard against crosses on ungrounded output leads. For example, if digit 1 is registered by operating relays (0) and (1), output lead Ll is grounded as required; and leads L4 and L5, though not grounded, are connected together. In many applications this is of no consequence but, if necessary, these crosses can be eliminated at the expense of more contacts. One possible solution is to provide an independent path consisting of two make-contacts in series for each output lead of the translating network. In Fig. 12-12 this would require the addition of two makes on the (1), (2), and ( 4) relays, making a total or twenty contacts, or four per relay, in the translating network.

12.9

SYMMETRIC

CHECKING

NETWORKS

The checking network for a self-checking code is usually some form of symmetric network. The use of a two-out-of-five symmetric network for checking the action of relays operating in a two-out-of-five code is illustrated in Fig. 12-12. This network is closed only for the ten combinations representing the digits. In the biquinary self-checking code a one-out-of-two circuit on the (00) and (5) relays placed in series with a one-out-of-five circuit on relays (0), (1), (2), (3), and (4) may be used. When some of the possible combinations which close a symmetric circuit are not used in a code, a symmetric circuit can often be modified to reject these combinations.

Codes and Translating

Circuits

299

As an example, certain applications may require that information in addition to the ten numerical digits be represented by two-at-a-time code combinations. A sixth element, designated "10", can be used with the usual 0, 1, 2, 4, and 7 elements to form five additional combinations. If some of the fifteen combinations are not used, the checking network may consist of a two-out-of -six symmetric circuit which is modified by omitting contacts so that the path is closed for only the desired combinations. Modified symmetric circuits which are closed for twelve and thirteen particular ones of the fifteen two-at-a-time combinations are shown in Fig. 12-13A and 12-13B. It should be recognized that the unmodified two-out-of-six symmetric circuit will also give a single error check in these examples, as will a circuit which indicates only that an even number of relays is operated. However, a circuit which is closed for only those combinations that are used gives a check against certain multiple-error conditions.

12.10 INTERMEDIATE

TRANSLATION

When there is a limited number of translating contacts on the relays or switch elements, it sometimes is not possible to translate directly to the desired output code. In these cases, auxiliary apparatus may be employed to perform an intermediate translation. An example of this is where digits are registered on the verticals of a crossbar switch having only two available make contacts at each crosspoint. The particular crosspoint (0 to 9) which is closed represents the value of the digit registered, and the two contacts of this crosspoint must ground the desired output leads. When the output code is one of the fourelement codes of Table 12-3, there is occasion for three output leads to be grounded. This can be done by means of an auxiliary relay as shown in Fig. 12-14. In effect, the digits are translated by the interconnections of the crossbar switch contacts into an intermediate five-lead code in which no more than two leads are grounded simultaneously. The auxiliary relay converts this to the desired four-lead code.

/Ed7··$P·"' -;-

-=-

OPEN FOR:

OPEN FOR:

10 ANO 2 10 ANO 4 10 ANO 7

A Fig. 12-13 Modlfled Symmetric

10 ANO 2 IOAND 7

B Check Circuits

The Design of Switching

300

Circuits

The method may be extended by adding a second relay as shown in dotted lines in the figure. The four output leads can then be grounded in any of the fifteen possible combinations by grounding either one or two of the resulting six input leads of the auxiliary translator. 12.11 TRANSLATION

BETWEEN

NUMBERING

SYSTEMS

In some translation problems a large number of input combinations must be translated to a corresponding number of output combinations with a systematic correspondence between the input and output combinations. An example is where a multidigit number in decimal notation is translated to an output code corresponding to some other numbering system or orderly method of designation. The procedure in designing these circuits is to determine the way in which the two systems correspond and use this as a clue in constructing the translating network. If is often possible to determine a systematic method of expressing output combinations in terms of input combinations. Equivalent contact networks are then easily developed. Suitable choice of the codes in which the original information is expressed will often lead to quite simple translating circuits. The principle a two-digit decimal

can be illustrated by a circuit which translates from number to a three-digit mixed-base number where

INPUT TRANSLATION

AUXILIARY TRANSLATION

2

2

9

OUTPUT BINARY CODE)

7 6

5 4

3

2

0

4

-

- -

8

4

4

_e;;;.... ___

..1....__

,.-If"I

II

~-o-Jl.-

•·a

·-t.-=-n

t._

' I : I

t __ .J

------11

L ____J 1

--~-

X-BAR SWITCH VERTICAL WITH 2 CONTACTS PER CROSS POINT

'-'

Fig. 12-14 Use of an Auxiliary Translator

a

Codes and Translating

301

Circuits TWO-DIGIT DECIMAL NUMBER 2ND DIGIT

1ST DIGIT (EVEN) 0 (000)

I

2

..

6

3

5

7

J

l l l l I.. 2

' C

5

I

8~

97

0

0

'

1ST DIGIT IBASE-51

3

4 (BELOW 5)

7

8

9 (SANO UP)

I I l l ..l

+

I 0 2 3 ~ 2ND DIGIT IBASE-4)

3

6

2

0

2

3

3RD DIGIT IBASE-51

THREE DIGIT MIXED BASE NUMBER

Table 12-10

INPUT (DECIMAL

Correspondence Between the Decimal Numbering System and a Mixed-Base Numbering System

0

(A2)

I

OUTPUT (MIXED BASEi

(A4) 2

8

t

--!---

4

4 FIRST DIGIT

FIRST DIGIT

3

D

2

D

I

Lqi

'.T: ♦

:}SECOND 2 DIGIT

3 0

(81) ♦

2

5

♦

(B4)

4 SECOND DIGIT

2

Fig. 12-15

--~=

0

Translation

THIRD DIGIT

3 4

D

D

from Decimal to a Mixed Base System

302

The Design of Switching

Circuits

bethe bases are 5, 4, and 5.• Table 12-10 shows the correspondence tween these two numbering systems. The first and third digits of the mixed-base number can be determined from the first and second digits respectively of the decimal number. The second digit of the mixed-base number is determined from both digits of the decimal number. As shown in the table, this digit depends on whether the first decimal digitis even or odd and on whether the second deci~al digit is or is not below five. If a choice of codes can be made, the 1-2 -4 -8 code is an obvious choice for the first digit and the 1-2-4-5 code for the second digit. The second (base-4) digit of the mixed-base number can then be determined from the "1" element of the 1-2-4-8 code and the "5" element of the 1-2-4-5 code. The circuit can be developed by conventional methods, and it is shown in Fig. 12-15 where the first digit relays are designated (A-) and the second digit relays (B-). Two factors contribute to the simplicity of this circuit. One factor is that the codes chosen for the first and second decimal digits fit into the mixed-base system. If some other codes are used, such as two-outof-five, for example, the indication of odd-even and below-five is more complex. The other factor is that there is a simple systematic correspondence between the two systems of enumeration. For example, a translation of a multidiglt decimal number to the base-2 system, or vice versa, is complicated because simple relations do not exist.

12.12 LARGE TRANSLATORS A major translating proble~ occurs in some switching applications when a large quantity of information must be translated to output information which has no systematic correspondence to the input information. An example occurs in large digital computers where frequently used mathematical formulae are assigned identifying index numbers. When one of these numbers is presented to the input of the translator the necessary information to permit the computer to solve this formula is delivered at the translator output. Another example is in the commoncontrol type of telephone switching system where the first three digits dialed by the subscriber, which indicate the office being called, are translated into instructions which permit the machinery to locate and establish a call over one of the trunks to this office. These translators act as "files" of information where the input signals, in effect, are arbitrary "index" numbers which identify various complex sets of infor~ation and permit any given set to be produced in response to its identifying index number. In general, the nature and • This is recognized as part of the system of designations of 10,000 terminals on the terparagraphs of minals of 500-point switches, which was mentioned in the Introductory this chapter.

Codes and Translating

Circuits

303 TRANSLATING DEVICES I RELAYS)

P.

CODE POINTS

_....e = --

SELECTING INPUT CODE DIGITS

NETWORK

\

@--

\

I

\

\

\

I

/

I 0--

'\ ,,

' ,\ \

,,____

v /'

.,'\

\

I

1 I

/

1

II \

I

lcRoss-coNNECTIONs

OUTPUT INFORMATION

I

I

\

.._

I llTt / \\ I J:--o-ll---: ~ ~t

I I I

\ I

'

Fig. 12-16

@...

0~ ----~

@--

e -----+_)

Basic Plan of a Large Translator

form of the input information bears no relation to the particular output information. To permit flexibility in their use, these translators are often arranged so that the output information corresponding to an input index number or "code" can be changed easily. The general principle of a large translator is to provide trans lating devices which can indicate to output circuits the complete output information corresponding to an input index code. The translator must then contain a means for selecting and activating a particular translating device when its assigned code is presented to the input. Large translators may employ relays, varistors, gas tubes, or special electromechanical devices. A typical arrangement of a relay-type translator is shown in Fig. 12-16. This consists of a selecting circuit which places ground on a code point corresponding to the input code. A code point is provided for every possible input code. The code points are crossconnected to translating relays which, in turn, have their contacts cross-connected to output terminals according to the desired output information. The output information may be in numerical form using one of the combination codes, or may merely be in the form of control signals which cause associated apparatus to perform specific actions. The cross-connections make this system of translation very flexible. Each translating relay can indicate any set of output information, and the code-point cross-connections permit this set of information to be associated with any one or more input codes. Physically the crossconnections are made between banks of wiring terminals and may be easily changed. The selecting circuit may take various are discussed in the following chapter.

forms,

several

of which

304

The Design of Switching PROBLEMS

FOR CHAPTER

Circuits

12

12-1

Design a simple translating circuit which will translate from the 1-2-4-8 code (0 = 2-8) to the 1-2-4-5 code (0 = 1-4). The input and output elements consist of leads on which the active condition Is indicated by ground. (This can be done with 41 contact springs.)

12-2

Design a simple translating circuit which will translate from the 1-2-4-8 code (0 = 2-8) to the 0-1-2-4-7 code. The Input and output elements consist of leads on which the active condition Is Indicated by ground. (This can be done with 53 contact springs.)

12-3

Design a simple translating circuit for the relays of the decade counting circuit of Fig. 11-19 (Chapter 11) which will deliver output indications by grounding five leads in the 0-1-2-4-7 code. (This can be done with 32 contact springs.)

12-4

Design a translating circuit which will translate from the 0-1-2-4-7 code to the 1-2-4-8 code. Provide a means for grounding an alarm lead (AL) if an error occurs In the input combination. (This can be done with 46 contact springs.)

12-5

Two base-ten digits are registered on ten relays using the additive two-out-of-five code. Design a translating circuit with two groups of output leads, the first group containing twenty leads and the second group containing five leads, such that one and only one lead Is grounded In each group for each combination of the two input digits. This provides a mixed-base output, with the first digit base-20, and the second, base-5. Keep the number of relays employed in the circuit to a minimum, and do not use more than 24 contact springs on any relay. (This can be done with 103 contact springs.)

12-6 The minutes of the hour from 00 to 59 are Indicated In a self-checking

code by a group of relays which operate continuously. The units digit (0 to 9) Is represented by five relays, UO, Ul, U2, U4, and U7 In the conventional two-out-of-five code. The tens digit (Oto 5) Is represented by four relays, TO, Tl, T2, and T4 In two-at-atime additive combinations where zero Is represented by T2 and T4 operated. Design a circuit which will light a red lamp (R) on the hour, the half-hour, and the quarterhours, (I.e. when the numbers Indicated are 00, 15, 30, and 45), and will light a green lamp (G) at five-minute Intervals within each quarter-hour (i.e. when the nwnbers Indicated are 5, 10, 20, 25, 35, 40, 50, and 55). In addition to the above, the circuit must light an alarm lamp (AL) at any time that an error occurs in the operation of the Indicating relays. (This can be done with 62 contact springs.)

12-7

What Is the minimum number of code elements required to provide binations In which the presence of single errors can be detected?

flfty

code com-

Asswnlng that each code element is represented by a relay, design a check circuit on the contacts of these relays which will light an alarm lamp when an error occurs. 12-8

The operation of certain teletype apparatus requires code combinations representto control the Ing the twenty-six letters of the alphabet and additional combinations operation of the printing typewriter as follows: 1. Shlft to Figures 2. Shift to Letters 3. Space 4. Carriage Return 5. Line Feed

Codes and Translating

305

Circuits

Outline the method or constructing capable of correcting single errors.

a code representing this information which ls Cover the following Items In the ouUlne:

1. 2. 3. 4.

Total number of code elements Number or Information elements Number of check elements Method of evaluating check elements for a given combination tion elements when constructing the code. (Give example) 5. Method or checking and correcting errors. (Give example) 6. Effect or double errors. (Give example)

How can the code be extended to detect double errors?

or Informa-

Chapter 13

cmcurrs

FOR SELECTING

A "selecting" circuit is specifically defined as a switching circuit arrangement which chooses from a number of similar items a particular one identified by predetermined information. There is no freedom of choice in selecting; the specified item must be picked, and no other can be used. In switching systems the many types of items from which selections are made include particular paths or terminals of an interconnection network, specific circuit units necessary to exert or accept control action, one of several information sources such as register circuits, and one of many code points in a large translating circuit. In practical selecting circuits, the selectable items are most often represented by single terminal points or single leads, although in some cases each individual item may comprise several terminals. The output of a selecting circuit is most often a control lead or leads by means of which the associated selectable circuit unit may be seized, controlled, or made busy. The act of "picking" in a selecting circuit usually involves marking the specified point or lead by applying to it ground or battery or other identifying circuit condition. The discussion in this chapter will be confined to selecting arrangements which make use of relays or similar two-valued devices. However, selecting can, of course, also be accomplished by other means such as a multipoint ·switch which ls directed by appropriate control to connect to a specified one of its output terminals. The chief characteristics of a selecting circuit are indicated in the block diagram of Fig. 13-1. This shows a signal (for example, ground) which is applied to a particular one of the selectable outputs according to the existing conditions on the control leads. For each output there exists a particular combination of control conditions, and a given output is selected by setting up the corresponding control combination. In effect, then, there must exist some "code" which associates control combinations with particular outputs. Thus, in its broadest aspect, a selecting circuit is a translating circuit which selects one of n outputs by translating from a combinational control code to a oneout-of-n output code. However, because of the magnib.lde of the usual - 306 -

Circuits for Selecting

307

,

~

... ,

SELECTABLE }

~

,

OUTPUTS

~

,

t t t CONTROL Fig. 13-1 The Basic Selecting Function

selecting problem, it is not practical developed in the preceding chapter. 13.1 ELEMENTARY

to handle it by the techniques

RELAY SELECTING CffiCUITS

The simple multiple of relay contacts shown in Fig. 13-2 is a very elementary selecting circuit. Grounding a particular control lead causes a connection to be established to the corresponding output. The restriction that only one control lead be grounded at any given time is implied. This is a trivial example of selecting, since the original control is on a one-out-of-n basis. However, in later sections of this will be used as a unit in assembling more chapter, this arrangement elaborate selecting circuits. A more important example of a simple selecting arrangement ls the transfer tree arrangement of relay contacts. An example is shown I I

I

SELECTABLE

OUTPUTS

I I

__L.

'---,.....--/ CONTROL Fig. 13-2

An Elementary

Selecting Clrcult

The Design of Switching Circuits

308

in Fig. 13-3 which uses three relays to select one of eight outputs according to the combination in which the relays are operated. A characteristic of transfer trees which is typical of more elaborate selecting arrangements described later is that the selection is accomplished in "stages." As can be seen from Fig. 13-3, each stage of a transfer tree consists of the contacts of a single relay, and the operation or release of this relay subdivides the possible choice of outputs until the final stage connects to the particular one.

NONE C

B

BC _______

A

SELECTABLE OUTPUTS

AC AB ABC

A

CONTROL

Fig. 13-3 Selecting by Means o( a Transfer

Tree

A property of transfer trees which is a result of the exclusive use of transfer spring arrangements is that all paths are disjunctive and the ipput (ground) can be connected to only one output terminal at any given ~ime. Although it is implicit in selecting circuits that one and only one output be activated at any given time, it is not always a characteristic of selecting circuits that all paths are disjunctive. For example, see Fig. 13-2. Since every input combination selects an output in a fully developed transfer tree, any sequence of action of the control relays when selecting a particular output may cause the selection to pass over several of the outputs momentarily. If this introduces objectionable consequences, it is necessary to hold open the output path until the desired combination of control relays has been established. Since the n relays of a transfer tree can make a selection among 2n outputs, it is the most efficient selecting arrangement possible in terms of numbers of relays required. However, due to the impracticability of constructing relays with large numbers of transfers, this type of circuit is not used in large selecting circuits.

Circuits

for Selecting

13-2 MULTIBRANCH

309 TREES: GENERAL

The transfer tree provides two branches toward the output at each junction point, or fork, of the circuit paths. The number of branches can be increased by using a multiple arrangement of the type shown in Fig. 13-2. With this arrangement, the contacts of all relays are makecontacts; and if each fork contains n branches, there must be n relays at each stage. As an example, a two- branch tree of three stages using only make-contacts is shown in Fig. 13-4. Each stage contains two relays, and one and only one relay in each stage is operated to make a selection. Fig. 13-4 is the make-contact equivalent of the three-stage transfer tree of Fig. 13-3. A two-stage three-branch tree would permit selecting one of nine outputs. It is not necessary that the number of relays (or branches) in each stage of a multibranch tree be the same. Fig. 13-5 shows a two-stage tree with a two-branch first stage and a four-branch second stage. A one-out-of-eight selection is made by operating one of the two first-stage relays, Ao and A 1 , and one of the four second-stage relays, B 0 , B 1 , B 2 , and B 3 . When the number of outputs is large, a selecting tree may be arranged in many ways. Optimum conditions in a particular case will depend both on theoretical considerations and on practical aspects

Ao •..._ __

___.[----4~-i--·-o-,--

I I

0

a

j---

I

I•

I ,____ _.__~_ -------o---

'---1-&--0

1-I

AoBo c, Ao B,

Co

A0 B1

c,

A, Bo Co A1 Bo C1

I

1

I

d}d} STAGE I

BoCo

STAGE 2

A, B,

t

~ STAGE 3

CONTROL

Fig. 13-4 A Three-Stage,

Two-Branch Selecting Tree

Co

A1 B1 C1

OUTPUTS

The Design of Switching

310

Circuits

Ao Bo Ao

Ao Bz

t

_r--

I I I

81

Ao 83 A1 Bo

f

A1 81

... "' ::,

... a Q.

I I I

Ao 82

I I I

I I

~ ~ ~ ~ ~ STAGE I

STAGE

Q]~ 83

2

Fig. 13-5 A Selecting Tree with a Two-Branch

and a Four-Branch

First Stage

Second Stage

having to do with available apparatus and control requirements. These several aspects of the selecting problem will be discussed separately.

13.3 THEORETICAL

CONSIDERATIONS OF MULTIBRANCH

TREES

A generalized diagram of a multistage multibranch selecting tree is shown in Fig. 13-6. From a consideration of this and the examples of Figs. 13-4 and 13-5, a number of statements relating to the size of these arrangements can be made. 1. If the number of stages is n, and the number of branches at R 1 , R 2 , ... Rn, then the maximum each stage is respectively number of outputs, M, is; R 1 x R 2 x ... Rn. 2. The number of relays in each stage is equal to the number of branches of that stage. Hence, the total number of relays in n stages is; R 1 + R 2 + ... Rn. 3. The number of paths into any stage is equal to the outputs of the preceding stage. Therefore the number of inputs to the kth stage is; R1 X R2 X ••• Rk-1 ·.

STAGE

2

---¥

---~}A

Fig. 13-6 General Multistage,

Multlbranch

1x A 2 x,·--,xAn

oo,eu,

Tree

L-:J,.:

L ____

-.....J,

TERMINAL CIRCUITS

Fig. 15-13 Gas-Tube Lockout Circuit wlth Oul!XJl Lead Isolation

lead, the element attempts to pass from its normal condition to its active condition through a negative resistance region in its characteristics. As already discussed, the negative-resistance characteristics in conjunction with the common inductive impedance must be such that only a single element may reach the active condition. This general solution to the lockout requirements may be used in the derivation of numerous electronic circuit arrangements for various specific applications. For example, in a particular instance it might be necessary that a gas-tube lockout circuit give positive-potential signals on the output leads as active indications, and zero potential as the normal indication. These requirements suggest that the output leads be connected to the main cathodes of the gas-tube lockout elements. However, to prevent momentary false indications on the output leads, the main cathodes should not be in the paths through which the tubes are fired. For this reason, tubes with separate control anode-to-cathode gaps might be employed, the ionization of a control gap transferring to the main anode-to-cathode gap. The negative resistance exhibited during this transfer could be utilized to obtain the lockout action. A lockoutcircuit operating in this manner is shown in Fig. 15-13. With all tubes extinguished, a positive potential is placed on one or more control leads by their respective terminal circuits. This potential is sufficient to break down the control gaps of the corresponding tubes.

Circuits

for Lockout

363

Lockout may or may not occur at this stage of operation, depending upon the characteristics of the tubes and the magnitudes of the circuit components. With the control gap ionized, the discharge transfers to the main - anode - to - starter - cathode gap. Lockout action is obtained during this transfer, since the main anodes and starter cathodes, respectively, are multipled together and the latter are connected to a common inductive impedance. Only one tube, then, can effect this transfer; and, as the transfer occurs, the multipled starter cathodes rise in potential sufficiently to extinguish all starter gaps. Finally, the discharge in the one tube remaining ionized is transferred to the mainanode-to-main-cathode gap, actuating the output lead connected to the main cathode of the tube. The potential on the output lead is maintained, and the tube remains ionized until the common battery supply lead is opened, restoring the circuit to normal. It is necessary in this circuit to maintain the main-anode-tostarter-cathode discharge, after transfer to the main cathode is completed. Otherwise, waiting terminal circuits could again fire their respective starter gaps when the starter cathodes returned to ground potential. Under these conditions, another tube could then be fired since there is no lockout action between main gaps. The resistance (Re) in the starter cathode circuit must be chosen, therefore, to allow sufficient current to enable transfer to the main gap and still maintain the starter-cathode-to-main-anode discharge. Individual resistors in the input control leads are utilized to limit the current on starter - gap breakdown to a low value.

PROBLEMS FOR CHAPTER

15

15-1 (a) The control leads of a ten-relay double-transfer lockout clrcult {Fig. 15-2D) are designated a0 through a9 • Lead aocontrols the relay having first preference ln the output transfer chain. Give the sequence ln which the output leads are grounded for the following sequence of grounding and opening or the control leads (overscore denotes opening). Some or the requests wlll be abandoned without obtaining service. Assume sufficient time between changes of control conditions for relays to act, and assume also that all the preference relays are identical ln characteristics. Control

sequence:

a 2 -a 8 -a 0 -a 7 -li2 -a 5-ll,-li: 8

a 1 -a5 -a 2 -a 9 -a0 -a 6 -1 1 -a 7 -ll 2 -ll 7 -ll 6 -a9 (b) Repeat part (a) for a gate-type lockout clrcult

(Flg.

15-5).

15-2 Four lndlvldual terminal clrcults compete ln a lockout circuit to establish a connection to a common circuit. It ls desired for a particular application that the terminal clrcults be served In a sequence closer to that In which they orlglnate requests than can be afforded by either the gate or the double-transfer lockouts described ln the text. A plan ls adopted, therefore, ln which the Hrst request to appear ls to be served Immediately, and the second request Is to establish a preference

The Design of Switching

364

Circuits

condition such that it will be served next, no matter what other requests may appear before the Hrst requesting circuit has released Its connection. Requests that are originated while a connection is maintained and after the preference (or the circuit to be served next has been established, are not necessarily served in the sequence in which they are originated.

A request (or service from a terminal circuit consists o( a ground condition corresponding control lead. The connection to be established from terminal to common circuit 'Consists o( a single lead. Design a relay lockout circuit nine relays.)

15-3

to fulfill

these requirements.

(This

on its circuit

can be done with

Ten primary terminal circuits, designated A 0 to Ag, are associated with ten corresponding secondary terminal circuits, designated B 0 to B 9 , through a lockout circuit. Upon request (a control lead grounded) from a particular primary terminal circuit to its circuit, a single-lead path is connected through from that primary corresponding secondary circuit, and all other primary circuits are prevented from obtaining connection to their corresponding secondary circuits. Only one primary circuit Is to be connected to Its respective secondary circuit at a time. A connection, once established, Is to be maintained until the primary circuit involved has terminated its request condition. Design the required relay circuit to perform (This can be done with ten relays.)

the lockout

and connecting

(unctions.

15-4 A gate lockout circuit (Fig. 15-5) is used to control the access from a number o( terminal circuits to a common circuit. As a means (or measuring the traffic handled by the lockout circuit, It is specHied that a ground pulse o( approximately forty milliseconds duration be supplied to an electro-mechanical counter over a single lead TM each time a terminal circuit obtains connection through the lockout to the common circuit. Design a relay arrangement controlled by the relays o( the gate lockout circuit which will provide the required pulses. Il may be assumed that the minimum duration o( each terminal-circuit-to-common-circuit connection is approximately 400 milliseconds. (This can be done with three additional relays.) 15-5 Design a lockout circuit to establish upon request an ungrounded one-wire connecto a single common circuit, meeting the tion from any one o( ten input circuits following requirements: (a) The Input request signal from each Input circuit Is ground on a single control lead. This ground is maintained until the connection ls to be opened. (b) When the lockout circuit is Idle, two or more Input requests originating simultaneously (within the acting time or a relay) constitute a condition which must light a common alarm lamp and prevent any connection as long as a double request persists. (c) Any requests occurring while an established connection ls maintained must be locked out and must light a common busy lamp which will stay lighted as long as the subsequent request persists. However, if the Hrst connection drops out while a single second request Is maintained, the second request shall establish a connection. U two or more requests are waiting when Vie first connection drops out, they shall be treated as requests originating simultaneously. (d) The circuit solution.

shall

employ

no resistors

(This can be done with eleven relays.)

or

shunting

techniques

ln its

Chapter 16 cmcuITS

FOR FINDING AND HUNTING

It is often necessary in switching systems to choose one item out of a group of similar items on some predetermined bas.is and to give an indication of the choice. The information derived from the circuit that performs the choice, known as a "locating" circuit, is used to determine which of the several items will be used in a particular situation, and most frequently serves to control a connecting circuit, either directly or indirectly. The broad class of locating circuits breaks down into three functional subdivisions: selecting circuits, finding circuits, and hunting circuits. From the circuit point of view, the selecting circuit, or circuit which picks a particular predetermined item from a group, forms a distinct subgroup which it is convenient to treat independently. This was done in Chapter 13, Circuits for Selecting. The other twofunctional subgroups, finding circuits and hunting circuits, are closely related circuitwise and are discussed in this chapter. A "finding" circuit can be defined as a circuit which locates from among a group of items one particular item which has itself initiated a request to be found. In general, the requests are introduced at random, and the finding circuit must act to give a positive indication of only one of perhaps several competing items. An example of the application of a finding circuit is in telephone systems where, when a subscriber initiates a call, his particular line must be found and identified. A "hunting" circuit, on the other hand, is a circuit which locates any available one of several identical items when it receives a request signal from an external circuit. Availability is usually based on the busy or idle condition of the individual items, with the choice restricted to idle items. A hunting circuit might well be used in a computer system to locate an idle register unit in a set of registers into which to place a number after an arithmetical operation. A block diagram indicating the relationship between a finding or hunting circuit and its environment is shown on Fig. 16-1. The blocks labelled G- represent the several items or circuits of the group from which the locating choice is to be made. Each G- block is associated with the locating circuit by means of a T- lead which, in the case of a - 365 -

The Design of Switching

366

Circuits

finding circuit, provides the individual request signal; and, in the case of a hunting circuit, gives a busy or idle indication. The information specifying the located G- item is provided by a signal on one of the output L- leads, which, as mentioned before, may actuate a connecting circuit to associate the corresponding G- item with a common circuit, or may simply make available the location information for use by some other functional circuit. The principal external difference between finding and hunting circuits lies in the origin of the start signal, carried the start signal is over lead ST of Fig. 16-1. With finding circuits, always provided by one of the G- circuits. Analysis of the situation indicates that the start signal can be applied to a common lead, as shown on the figure. This lead may control a connector to associate the finding circuit with the G- circuits, or it may initiate appropriate action within the finding circuit. If the finding circuit is permanently connected to the G- circuits, a signal on one of the T- leads may be sufficient to start action without the necessity of a separate ST lead. If the locating circuit is used for hunting, the start signal always originates in some external circuit which desires the locating of one of the G- circuits. If the hunting circuit is a common circuit, the ST lead may control a connector to associate the hunting circuit with the desired group of G- circuits; if the hunting circuit and G- circuits are permanently connected, the ST lead may supply ground to a contact network on the hunting relays in order to give an output indication. 16 .1 BASIC REQUffiEMENTS

OF FINDING AND HUNTING

CIRCUITS

From a careful analysis of the circumstances under which finding and hunting circuits must work, certain basic requirements can be deduced. These are in addition to considerations affecting the start control and the method of connecting the locating circuit to the associated

HUNTING ---•ST FINDING

}

LOCATION IDENTIFICATION

}

TEST OR REQUEST LEADS

}

GROUP OF SIMILAR CIRCUITS

LOCATING CIRCUIT ST

I FINDING OR HUNTING)

Fig. 16-1 Relationships of Circuits

and Leads In Hunting or Finding

Circuits for Finding and Hunting group of circuits. design problem. considerations:

367

Factors such as these latter tend to vary with every The basic requirements derive from the following

Finding Circuits (1) Circuits that may request finding action can, in general, initiate requests independently and at any time. (2) If a request is entered while a previous request is being served, the second request must not be able to interfere with the established output condition of the finding circuit. (3) Since requests may, in general, be entered at random, the finding circuit must be able to accommodate simultaneous requests and give an output indication corresponding to only one request at a time. Hunting Circuits (1) The circuits of the group in which the hunting action is taking place are available (idle) or unavailable (busy) independently of each other. A circuit may change from a busy to an idle state at any time. In the usual situation, the hunting circuit is the only circuit with access to the group, and individual group circuits cannot become busy except under control of the hunting circuit itself. However, in multifunctional circuits, which are not discussed here, it is sometimes necessary to permit two or more hunting circuits to work in a single group simultaneously. Under such circumstances, special circuit means are required to guard against adverse effects from group circuits becoming busy from other points during a hunting action. (2) If, after the cuit of a group, the second idle output condition

hunting circuit has located a particular idle ciranother circuit becomes idle, the appearance of circuit must not interfere with the established of the hunting circuit.

(3) When the hunting circuit starts its hunting action, several circuits in the associated group may be idle concurrently. The hunting circuit must be able to give an output indication corresponding to only one of the available circuits at a time. The analysis of the above considerations leads rapidly to the conclusion that the basic requirements of hunting and finding circuits are identical with the basic requirements for lockout circuits. Therefore, although the circuit concepts are functionally different, in circuit form hunting-finding circuits are identical or similar to the lockout circuits discussed in Chapter 15. This is clearly indicated in Fig. 16-2, where

The Design of Switching Circuits

368

IDENTIFICATION

OF LOCATED CIRCUIT

ICON)

T.!l. TEST OR REQUEST LEADS FROM LOCATION GROUP

Fig. 16-2 Typical Locating Circuit When Request Signal (Finding) or Idle Indication (Hunting) Is Ground on T- Lead

the configuration of a double-transfer lockout circuit is used for either finding or hunting in a situation where the finding request signal or the idle hunting indication is an active (grounded) condition on one of the Tleads. If the circuit of Fig. 16-2 is used for finding, the common ST lead is grounded by any one or a combination of the group of circuits which may request finding action. In addition, any requesting circuit concurrently grounds its T- lead. The ST lead operates a connector relay to associate the locating relays with the T- leads, and one or more of the (L-) relays operates. An output indication corresponding to the operated (L-) relay in highest preference appears on one of the L- leads. If simultaneous requests are present, or if subsequent requests appear, they cannot affect the output until the original winning request is withdrawn.

When the circuit of Fig. 16-2 is used for hunting, the ST lead is grounded by an external circuit which requires the locating of an available circuit in a group of identical circuits. With circuit availability indicated by ground on the associated T- lead, the chosen output will again be supplied by the operated (Ll) relay in highest preference without possibility of interference from other circuits in the group. In lockout circuit applications, it is usually convenient to enter the lockout with an active condition on the service request lead. However, in hunting-finding circuits (particularly hunting circuits), circumstances may dictate the removal of an active condition of the T- lead as

Circuits

for Finding and Hunting IDENTIFICATION

369 OF LOCATED CIRCUIT

lCON)

1r~1 J= -=-

ST TEST OR REQUEST LEADS FROM LOCATION GROUP

Fig. 16-3 Typical Locating Circuit When Request Signal (Finding) or Idle Indication (Hunting) Is Removal of Ground from T- Lead

the request signal or the idle indication. This requires a change in the basic circuit to prevent the release rather than the operation of a higher preference locating relay during the hunting or finding process. This is easily accomplished by rei:ilacing the lockout transfer chain of Fig. 16-2 (which prevents higher preference relays from operating) with the lockup contacts of Fig. 16-3 (which prevent higher preference reIn addition, a timing relay is necessary to hold lays from releasing). open the output signal until the locating relays have established their condition. Certain other basic requirements for hunting and finding circuits develop logically from their functional application. For example, when a circuit is located after entering a finding request, external circuit action is initiated which will eventually result in the withdrawal of the original request signal. In similar fashion, when an idle circuit is located in a hunting action, external circuit action will eventually cause the original circuit to be made busy. In both these cases, the change in condition from request to no-request or from idle to busy of a located circuit must not be allowed to affect adversely the subsequent action of the locating circuit. Most frequently, this requirement is met by coordinating the sequential action of the locating circuit with respect to the other functional circuits comprising the multifunctional unit, rather than by modifying the locating circuit itself. However, this must always be kept in mind as one of the crucial factors in the design and application of finding or hunting circuits.

370 16.2

The Design of Switching

Circuits

FINDING CIRCUITS

By their fundamental nature, finding circuits most closely approach lockout circuits in requirements and function. Unless unusual circumstances intervene, the normal input request signal to a finding circuit is an active condition, and therefore a standard lockout configuration is satisfactory. In general, any of the circuits of Chapter 15, including double-transfer, gate, end-relay, multiple-stage, and electronic lockouts, are suitable for finding. The specific type used will depend upon the requirements of the particular application. The use of a common start lead and connector arrangement may or may not be necessary. 16.3

HUNTING CIRCUITS

Hunting circuits tend to vary more in their subsidiary requirements than finding circuits. In addition, the idle-busy indication is frequently the inverse of the usual lockout input signal so that a less familiar circuit form is required. For both these reasons, several circuits, designed to meet different requirements, will be presented in this section. The first circuit to be designed must meet the following ments in addition to the basic hunting requirements. (1) The circuit must respond to ground no ground as an idle indication.

require-

as a busy signal,

and to

(2) The location indication must be returned on the same T- lead that is used for the individual busy-idle indication. (It may be assumed that the terminus of this lead makes the corresponding circuit busy in addition to indicating its status. (3) In addition to supplying the location indication, the hunting circuit must close through a group of several leads corresponding to the located circuit.

(4) A relay· (HS) must be operated successful.

to indicate

that

(5) A relay (AB) must be operated the hunt-group are busy.

to indicate

that all circuits

the hunt was in

Consideration of the first three requirements indicates that the final circuit can most conveniently be designed with two relays per input. For example, requirements (1) and (2) imply that a relay per input must first recognize absence of ground on a lead as an idle indication, and then must be able to apply an output ground back on the same

Circuits

for Finding

and Hunting

371

(HS)

l•q,_

(CON)

tp T

.

[ Fig. 16-4

I,_ Hunting Circuit:

.r ""

Ground as Busy Indication

lead. This is difficult to achieve in a straightforward fashion with a single relay per input without loading up the relays with alarge number of control contacts. However, with two relays per input, one set of relays can indicate initially by their operated or released condition the busy or idle state, respectively, of the associated circuits. A transfer chain on these relays can operate on a preferential basis an auxiliary hunting relay corresponding to the first idle circuit. By utilizing a transfer chain on each auxiliary relay to open a common operating path and concurrently close its own locking path, the circuit can protect itself against subsequent changes of condition on the input test leads. The circuit is shown on Fig. 16-4. This configuration permits the operated auxiliary hunting relay, (HA-), to return a ground signal on the associated T- lead without any difficulty and by means of a set of make contacts, to close the paths specified in requirement (3). Requirement (4), that a separate relay, (HS), must be operated to indicate that the hunt was successful, is easily met by a relay in series with all (HA-) relays. Although operation of (HS) could be accomplished just as well by paralleled make contacts on the (HA-) relays, this method saves contacts. Requirement (5), for an all-busy relay facility. When all circuits in the hunting group the (H-) relay transfer chain can operate an will lock and open the operating ground supply

(AB), is met with equal are busy, a path through (AB) relay, which in turn for the (HA-) relays.

The Design of Switching Circuits

372

The general type of circuit just described can also be adapted to a hunting circuit where ground on a T- lead is an idle indication. Additional requirements on the circuit might be: (1) The indication of a located circuit must be given over an Llead independent of the input T- lead. As soon as it has identified an idle circuit, the hunting circuit must permit an immediate change from idle to busy indication on the T- lead corresponding to the located circuit. This requirement must often be met when the hunting circuit output directly shifts the chosen circuit from idle to a busy condition. (2) The circuit must permit a systematic rotation of preference. This requirement is often imposed in order to equalize the usage of a group of circuits. The first change required in the basic circuit of Fig. 16-4 to meet these requirements is the inversion of the (H-) relay transfer chain so that an operated instead of an unoperated (H-) relay operates an auxiliary hunting relay. This eliminates the need for the timing relay of Fig. 16-4, which was necessary to cover the operating time of the (H-) relays. In order to meet requirement (1), an adaptation of the gate principle. used in lockout circuits may be utilized. That is, as soon as a Lo

L-

Ln

(CONl FROM PREFERENCE SEOUENCE CIRCUIT ST

l

Fig. 16-5 Hunting Circuit:

l_ Ground as Idle Indication

Circuits

r----o...._

for Finding

and Hunting

373

IABl

rMftl

- - -

ILn

~L...o

-,~·+I

To

ST

T*

1B1-!CIRCUIT

IBJ,½ G0

CIRCUIT

Tn

* G-

*

(BJ,½ CIRCUIT

Gn

*CONTACT (Bl MUST NOT CLOSE UNTIL ACTIVATING GROUND HAS BEEN REMOVED FROM ST LEAD

Fig.

16-6

Hunting Circuit

Permanently

hunting relay has locked in to identify to open all test leads to obviate effects dition. A circuit with these changes is (CON) relay serves the double function The final requirement, plished by the usual method in a ring and tapping on to circuit can then advance the

Associated with Hunting Group

a located circuit, the gate can act from subsequent changes in conshown in Fig. 16-5. Note that the of connector and gate.

that of rotating preference, can be accomof connecting the preference transfer chain it at several po!nts. An external sequence preference on any desired basis.

Circumstances sometimes make it desirable to associate a hunting circuit permanently with its hunting group rather than through a connector. Reasons for this may be that the hunting circuit has onl!' one group with which to work, or that maximum speed of operation is required. In this case, the relays of the hunting circuit continuously follow the conclition of their associated circuits, and the start signal merely applies ground to an output contact network on the hunting relays. Again it is necessary, during the hunting process, to prevent relays from changing condition. This latter requirement could most simply be met if the start lead input were able to supply the output location signal and lock up operated relays concurrently. This can be done by utilizing secondary windings on the hunting relays to prevent back-ups, as shown on Fig. 16-6. With this circuit, at the moment the start signal appears, the hunting relays are in some particular configuration corresponding to the state of the circuits of the hunting group. [ Note that an operated (H-) relay indicates busy] . The ground from the start lead travels through the transfer chain to provide an output at the first unoperated (H-) relay. At the same time, the start ground holds operated all relays higher in preference than the output relay.

The Design of Switching

374

Circuits

A fundamental hazard in this scheme as described is that, if all hunting relays are operated when the start lead is grounded, the circuit will block. This requires provision of the (AB) relay to open the start lead until at least one (H-) relay is released. In application of the circuit of Fig. 16-6, care must be taken to co-ordinate properly the closure of the (B) contact in the located circuit and the removal of ground from the start lead. If the (B) contact is closed early, the start ground will be advanced to a second L- output. This sequential co-ordination must be accomplished by means external to the hunting circuit. When the hunting circuit is permanently associated with the huntgroup, the equal distribution of usage to the latter circuits can be achieved without difficulty. A logical method of doing this is to lock operated each hunting relay when its corresponding circuit is located on a circuit usage until all circuits in the hunt-group have been assigned. Then the common locking path is opened, permitting the hunting relays of any idle circuit to release. This requires a timing circuit to control the locking path to insure that a sufficient interval is provided for all idle relays to release. The circuit, with the locking and timing portions drawn with heavy lines, is shown on Fig. 16-7. This circuit also requires careful co-ordination between removal of the start signal and closure of the (B) contact in a located circuit to insure proper operation. A requirement may sometimes be imposed that a permanently associated hunting circuit must be able to accept an immediate return of busy over the T- lead from a located circuit, before it is possible to remove ground from the start lead. As mentioned previously, this ST

--

To

*r;:;-:i.

O,a,Lo

CIRCUIT G0

T_

Tn

*~

*r:::-:i..

O,a~ L-

CIRCUIT

G-

O,a~ Ln

CIRCUIT

Gn

*CONTACT 18) MUST NOT CLOSE UNTIL ACTIVATING GROUND HAS BEEN REMOVED FROM ST LEAD

Flg. 18-7 Permanently Associated Hunting Cl~cult which Insures

Equal Service

Circuits

for Finding

)f....__L:TCK

.__>---0----L

and Hunting

375

START RELAY

Ln

Lo

Fig. I 6-8

requirement The problem

Permanently

Associated

Hunting C Ir cult with Double Lockout Chains

is not met by the circuit of either Fig. 16-6 nor Fig. 16-7. resolves itself into these two phases:

(1) Return of the busy signal on the T- lead of a circuit just located must not advance the start ground to a lower preference output. (2) Return to availability of a higher preference circuit during the hunting process must not shift the output of the hunting circuit. Statement of the requirements in this way suggests combining a double-transfer lockout circuit with a set of hunting relays as shown on Fig. 16-8. Although the hunting relays can change their condition at any time while the start lead is grounded, the characteristics of the lockout circuit permit only one locating output lead to be activated. The circuit requires a separate locking path for the lockout relays to prevent release when the operating path advances.

16.4

ELECTRONIC

HUNTING

CffiCUITS

The inherent lockout characteristics of gas tubes make this device eminently satisfactory for performing the hunting function when speed is an important consideration. Any of the gas-tube lockout circuits discussed in Chapter 15 is adaptable to hunting, and a typical circuit is an idle circuit is indicated by high shown in Fig. 16-9. In this circuit, positive voltage on the T- lead. When the connector is closed to associate the hunting circuit with the individual circuits, tubes connected to idle circuits attempt to fire across the control gap. The lockout

The Design of Switching

376

2200n

r-

r-

I I I

I I I

1

1

_l_

= 1oovlL.I_

--------, CIRCUIT G-PATH

CLOSED FOR IDLE

I I

L_ To

_...,._

(Lo)

_

,-1 I

I

I

I I

I I

I I

-~~~l!_l~~J T_

____.,__--+-

-------, CIRCUIT Gn I I I

I

I

LIL-l

___

Circuits

-PATH

CLOSED FOR IDLE

I

I

-~~~~o~J Tn

!Lnl

~-----+--~11~

CONNECTOR

2000n

I MEG

I MEG

8200.0.

-:

Flg. 16-9 Gas-Tube Hunting Circuit

action imposed by the common impedance connected to all control cathodes permits only one tube to break down and transfer to the main gap. If the current through the control gap is maintained by delaying the shift of the located circuit test lead from the idle indication to busy, the voltage drop in the common anode-cathode resistors will prevent any other tube from firing until the connector releases and reoperates. In the circuit as shown in Fig. 16-9, the output indication is obtained from the operated (L-) relay in the main gap circuit of the conducting as an output, it can be obtube. If a potential change is satisfactory tained directly from a load resistor in the main gap anode or cathode circuit. The gas-tube hunting circuit has no fixed preference characteristics as had the corresponding relay circuit. Instead, the preference depends upon the indeterminate ionization and transfer-time characteristics of the individual tubes comprising the hunting circuit. If this is unsatisfactory in a particular application, it is possible to impose an effective preference upon a gas-tube circuit by placing R-C networks with graduated delay in the input paths to the control anodes. Because of the similarity between the lockout and hunting functions, any electronic devices exhibiting negative-resistance characteristics that are suitable for lockout purposes can be used in hunting circuits. Thus, the principles discussed in Section 15.6 (Chapter 15) under the heading "Electronic Lockout Circuits" are directly applicable to the design of electronic hunting circuits.

Circuits

for Finding and Hunting

377

PROBLEMS FOR CHAPTER 16 16-1 The figure below indicates a connector and control arrangement which give one hundred T- circuits access to five C- circuits. The T- circuits are divided into subgroups of ten, with each subgroup associated with its own connector and control. Within each connector, a fifty-lead single-link connecting path is satisfactory.

SUBGROUP ( OF 10

I I

Too

I

I Co CONNECTOR

To9

I

I I

I CONTROL

I

SUBGROUP ( OF 10

I I

T90

~

I

I

I

/--

TO OTHER CONNECTORS

BUSY-IDL E INDICATION (SINGL E LEAD PEfl C- CIRCUIT)

CONNECTOR

T99

I

C5

I I

I

I

1 CONTROL

I

Design the connector and control occur at random, and the control nectors and C- circuits without multifunctional circuit which can

~

circuit. Connector requests from the T- circuits circuits must permit maximum usage of the con-. any possibility of double connections. (This is a be done with thirty-five relays per connector.

16-2 A group of twenty telephone lines are served by a set of ten central office paths, any one of which may be used to complete a connection. Each of the ten central office paths terminates on a non-homing, back-acting 22-point rotary finding switch, whose bank terminals are wired to the telephone lines. Associated with each telephone line is a line one of these two relays operates, one finder telephone line through to the central office line relay releases. The other relay may be

circuit which contains two relays. When switch locates the line and connects the path. The connection is held until the used for control.

Finder switches are assigned to calling lines by a single hunting switch of the same type as the finder switches. The hunting switch should normally assign an idle finder switch prior to receiving a request, although it should be prevented from hunting when all finders are busy. Design circuits for the line, finder, and hunting circuits. The switches may contain up to six arcs (of which two are reserved for the talking path) and supplementary relays may be used. (This can be done with one relay per finder switch.)

Chapter 17 CffiCUITS FOR TIMING There are innumerable situations in switching design that callfor circuits or apparatus to produce timed intervals. These situations arise from a variety of causes which in practice are either stated in the requirements of the problem or may easily be recognized by the circuit designer. Many cases of timing in switching systems occur when it is necessary to generate a specific time interval against which a circuitacting period may be measured. This may be done for trouble detecting, for example. If the circuit action is not completed within the predetermined interval, some alternative action is taken automatically, or an alarm is given, depending upon the specific circumstances. Other situations requiring timing are engendered by circuit considerations or the characteristics of apparatus. Race conditions between groups of relays often require specific timed intervals to insure correct circuit operation. The crossbar switch is a two-step device which requires a definite interval between operation of a select magnet and operation of a hold magnet. When a circuit tests for a condition which may or may not be present, it is often necessary to provide a definite minimum time for recognition of the condition. These siblations frequently can be handled by providing an appropriate sequence of operation instead of generating specific timed intervals. In these cases, the timing function is embodied in the design but does not appear as such. Another class of timing makes use of circuits which generate recurrent pulse intervals, usually for signaling purposes. These pulsing circuits are discussed in detail in the next chapter. Since the requirements for timing cover a wide range, both for over -all time and for accuracy, a variety of circuit and apparatus techniques must be used. In some siblations a simple device such as a slow-acting relay is adequate. In others, circuits of considerable complexity are necessary. In each particular design problem requiring timing, the factors of over-all time, accuracy, and available control means must be weighed in choosing the type of timing circuit or device. In this chapter, several important timing methods in current use are discussed, together with their relative advantages and limitations. - 378 -

Circuits for Timing

379

Attention is confined to the generation of timed intervals ranging upward from the millisecond range; circuits for timing extremely short intervals are not included.

17.1 STANDARD SLOW-ACTING RELAYS• The most frequently used timing device in relay switching circuits is the slow-acting relay. The need for this device is so widespread that slow-acting types with a great variety of winding and spring combinations are provided within the framework of commercial generalpurpose relays. The slow action is achieved by means of copper or aluminum sleeves around the relay core, and in some cases by special construction of the magnetic structure of the relay. The circuit requirements for relays which must provide timed intervals vary greatly. In some cases, where it is necessary to cover the operation of fast-acting relays by a timing relay, the operating time of an ordinary general-purpose relay with a heavy spring load and a high-time-constant winding may be sufficient. In this case the timing relay would not necessarily appear on the circuit drawing with an indication that slow action was involved. However, in the majority of situations requiring relay timing it is necessary to choose a relay definitely designed for slow action. Relays with aheavy sleeve can be obtained with minimum operating times up to the order of 100 milliseconds. Particular combinations of winding, sleeves, springs, and adjustments will give lower minimum operate times up to this value, so that the designer has considerable leeway in choosing an appropriate relay. It must be remembered, however, that in choosing a slow-operate relay to cover a certain time interval, the relay which gives a desired minimum time is also subject to a much greater maximum operate time. The spread between maximum and minimum times for relays of a particular type is often of the order of three-to-one, so that to obtain a guaranteed time interval of 60 milliseconds, the designer is usually forced to accept a possible time of 180 milliseconds unless special timing adjustments can be specified. This may or may not be of importance in the operation of a particular circuit, but it must always be kept in mind. In certain cases where over-all circuit holding time is of the utmost importance, this large spread makes it questionable to use ordinary slow relays for timing purposes. • This section applies only to general-purpose relays Including a conducting sleeve or slug on the core to Increase acting time. Specially constructed slow-acting relays are not dlscussed.

The Design of Switching Circuits

380

A simple illustration of the use of a slow-operate relay for timing is shown in Fig. 17-lA. When relay (A) operates to close output (1), there is a delay equal to the operate time of relay (T) before output (2) is closed. In practice, the release time of general-purpose be made considerably longer than the operate time.

type relays can

Relays designed specifically for slow release carry the possible minimum release time up to the order of 500 milliseconds. Again, there is a comparatively large spread between minimum and maximum times for a particular relay type. If time intervals of this order are required, a slow-release relay can usually be employed. Use of a slow-release relay, of course, implies that the relay is operated well ahead of the time when its release characteristics are utilized in the circuit sequence. This is illustrated in Fig. 17-1B .

.r----..._

______

A~· r

lr-•"

------------111 SLOW OPERATE RELAY

A

r---¾

~

f~

,_nJ}-l•q_

m ti~

1• 1

----·--L-.

r-1-----------111 =

,21

SLOWRELEASE RELAY

B

_c--

?7M-:i.=, rL-.,2, ----------------•► Ill SLOW RELEASE RELAYS IN TANDEM

C Flg. 17-1 Use of Slow-Acting Relays for Timing

Circuits for Timing

381

If a longer time interval is required than is possible with the operate or release time of a single relay, relays in tandem can be employed. This is shown in Fig. 17-lC, which uses three pre-operated slow-release relays. The relays can be both operated and released in tandem, or, if relatively fast pre-operation is desired, the relays can be operated simultaneously and released in tandem. The utility of this principle can be extended by employing other cycles of slow - acting relays.

The acting time of relays can be appreciably affected by external circuit elements. A resistor in series with the winding of a relay may be used to increase the operating time by reducing the margin of circuit current over the just-operate current. A low resistance bridged across the relay winding increases the release time by permitting the magnetic flux in the relay structure to decay slowly after the operate or hold path is opened. The same effect can be achieved by bridging across the winding a varistor poled to oppose the operating current. Another circuit element useful in delaying relay operation is the thermistor. This device has the property of increasing its conductance tremendously by thermal effect after current is permitted to flow through it during a period of time. A properly chosen thermistor in series with a relay winding will hold the current below the non-operate value for an interval after circuit closure. When the thermistor has heated to the critical value, the current will increase to the point where the relay can operate. Two disadvantages of this technique are that the variations in time are large, and that the cooling period required between successive operations is comparatively long. Obviously, the use of slow-acting relays can be extended to generate almost any desired time interval. It is usually uneconomical, however, to use more than two or three relays for the sole purpose of timing. The major disadvantage of slow relays is the large spread between maximum and minimum times, although in many applications this may be of little consequence.

17.2 INTERRUPTER TIMING CIRCUITS An important and much used method for obtaining time intervals, particularly those involving several seconds, is to employ a counting circuit actuated by a mechanically or electrically driven interrupter. The most frequently used spring combinations on the interrupters are simple makes and break - make transfers. The time per interrupter cycle may range from a fraction .of a second to hundreds of seconds.

The Design of Switching Circuits

382

---I

CYCLE INTERRUPTER CYCLE

,------7 I

I L---

F

I

8 _.J

MIN. TIME • (C-l)(W+X+Y+Z) MAX.TIME•

+ )(

C(W+X+Y+Z)+X

WHERE C • NO, OF CYCLES COUNTED

Fig. 17-2 Timing Circuit Actuated by Transfer

Interrupter

The technique in using the interrupters consists of providing a counting circuit which will count any desired number of cycles. A typical circuit to count two cycles of pulses from an interrupter with a transfer-type spring combination is shown in Fig. 17-2. The counting action begins when lead "TM" is grounded. It can be seen by analysis that with this type of circuit, designed to accept the ground closures for C interrupter cycles, the minimum and maximum elapsed times before the last relay operates are: Minimum time = (C-1) (W + X + Y + Z) + X Maximum time = C (W + X + Y + Z) + X where, W, X, Y, and Z are the time intervals of the individual parts of the interrupter cycle. With this type of circuit each pair of relays does not, strictly speaking, time a complete cycle, but rather responds to the beginning of each ground closure within the cycle. However, if period X (the open period following the first effective ground closure) occupies the greater part of the cycle, the approximate effect is that of timing a complete cycle with each pair of relays. If a circuit for counting pulses from an interrupter. with a single make-contact is employed, the minimum and maximum times involved in counting the ground closures of C cycles are:

Minimum time Maximum time

= (C-1) (W + X) + X =C (W + X) + X

where W and X represent the time duration of the closed and open parts of the cycle, respectively.

Circuits for Timing

383

With both these circuits, the uncertainty in the length of the timed interval is equal in length to the time of one complete cycle. The per cent variation in total time measured depends upon the number of cycles counted and the manner in which the time is allotted to the intervals within the cycle. For instance, in either type of circuit, if the W period occupies the major part of the time within the cycle, and if the ground closures of two cycles are counted, the total variation is of the order of two to one. If X is much greater than W, the variation for two cycles is three to two. Furthermore, as the number of cycles increases, the variation decreases in terms of the ratio of maximum to minimum time. In absolute terms, of course, the time variation is equal to one cycle. Both these circuits can be used to measure a single cycle of the interrupter if the X period is made the larger and more important part of the cycle. If X is the smaller part of the cycle, the variation in total measured time for one cycle becomes excessive. In general, the counting circuit used as part of the timing circuit may be any of the types discussed in Chapter 11. In conclusion, it can be seen that the interrupter timing circuit is well adapted to measuring long time intervals, if variations of the time of one cycle can be tolerated. The absolute variation of time can always be reduced to whatever value is considered necessary by decreasing the cycle length and increasing the count. 17.3 CAPACITOR-TIMED

RELAYS

When a circuit situation arises where it is necessary to generate a time delay up to a few tenths of a second within relatively close limits (of the order of tlO per cent), the capacitor-timed relay circuit may be used. This circuit consists of an electrically biased polar relay whose operation is controlled by the charge or discharge current of a capacitor. By holding the circuit variables to close limits, and choosing the operating point at an appropriate location on the R-C exponential curve, the variation between minimum and maximum operating times can be closely controlled. The basic theory of this circuit embodies the relations involved in the charge and discharge of a capacitor through a series resistance. Including a relay winding in the circuit introduces inductance which modifies the current-voltage relations, but with suitable polar relays the inductance of the windings is minor compared to the circuit resistance and can generally be neglected. Also, mutual inductance between windings is negligible. Therefore the circuit can be treated simply as resistance and capacity.

The Design of Switching Circuits

384

In Fig. 17-3 is shown a simple R-C circuit with means for applying either a potential E or a short circuit across R and C. When the switch is operated from position a to position b, the variation of current with time is shown as curve A in the figure, and is expressed mathematically as: l• -- E _,.., ..--t/RC

(17-1)•

R

The initial current is solely a function of E and R, and the rate of decay is determined by the product RC. If, after sufficient time to permit essentially full charge on the capacitor, the switch is thrown from b to a, the current flows in the reverse direction and its decay is again represented by curve A.

The voltage across the resistor represented by curve A where eR

during the charging

( 17-2)*

= E-~-t/RC

During the discharge period eR is reversed

period is also

in polarity.

'Since the voltage across the capacitor is the difference between the applied potential and eR, it follows curve B during the charging time and is equal to ec = E (1 - ct/RC)

(17-3)•

However, during discharge, the capacitor voltage decays from E to zero so that ec is represented by curve A and is equal to ec = E,ct/RC

(17-4)t

It should be noted that, if the voltage across the circuit is reversed, after charging the capacitor, the effect is that of doubling the voltage, since the circuit voltage and the capacitor voltage add together. This can be utilized to increase the effective voltage, where circuit considerations make this desirable. • If a residual voltage ex exlsts on the capacitor at tlme t = O when the swltch ls operated to position b, the current and voltage relations are modified to the more general form glven below. Note that posltlve ex alds battery potential E.

(17-2)

l = ~-e-t/RC R er = (E + ex)•ct/RC

(17-3)

ec = E - (E + Ex)• c 1/RC

(17-1)

t

If the discharge cycle (switch b to a) ls started before the capacitor charge, the effective capacitor voltage ls ex Instead of E.

(17-4)

ec = ex•ct/RC

has reached full

Circuits for Timing

385

(tl OR (fl

EXPONENTIAL

=

CURRENT OR VOLTAGE CURVES

!e-t/RC

(CURVE Al

I =-j,--t/RC

(CURVE Al

•R=

Ee-t/RC

(CURVE Al

eR=-E,.-t/RC

(CURVE Al

•c =

E(1-e-t/RC)

•c =

(CURVE Al

Fig. 17-3

(CURVE Bl

Current-Voltage

E ,--t/RC

Relations ln R-C Circuits

Since in capacitor-timed relay circuits the capacitor charge or discharge current is used directly, curve A and equation (17-1) will be most useful in the discussion. The simplest and perhaps most useful circuit for obtaining a delayed operate timing action is shown in Fig. 17-4. With the control relay normal and battery-ground connected, the current flow through the two windings holds the polar relay on its back contact. The primary winding is poled in such a way as to attempt to operate the relay, but the secondary winding, poled oppositely, overcomes the primary by a considerable margin to hold the relay non-operated. The capacitor, C, ls completely discharged at this time. The relay constants shown are for a typical sensitive polarized relay. When the control relay operates, current through the secondary winding immediately starts charging the capacitor. This initial charging

The Design of Switching Circuits

386

--ti> tr--

CONTROL RELAY

~ -=- pt

s

+

Rp

R1

=

=

PRIMARY 200 .Cl ± 1°4; EXACTLY SECONDARY 2600 fl. :!: I 04; EXACTLY TEST WINDING

TEST FOR

p

OPR. NON-OPR.

p

* 6 AMPERE ** 1.8 AMPERE

TEST IN MA. 3* 0.9**

200+rp 2600+rs

2000 14,000

TURNS TURNS

READJUST. IN MA. 2.8 1.0

TURNS TURNS

Fig, 17-4 Capacitor-Timed

Relay Circuit

current must be great enough to hold the relay on the back contact. However, as the capacitor charges, the secondary current decreases, value following curve A of Fig. 17-3, until the secondary ampere-turn drops below the opposing primary ampere - turn value by an amount sufficient to cause the relay to operate. This is illustrated in Fig. 17-5, where relay ampere turns are plotted against time. The second~ry ampere turns (Ns is) are initially much higher than the primary ampere turns (NpIp), as indicated by the point where the solid curve crosses the zero time axis. When the control relay opens the short circuit across the capacitor, Nsis starts decreasing at an exponential rate determined by RC. The relay will operate when Ns is decreases to the point where it is equal to Nplp - Ni, Ni being the ampere-turn value at which the relay is adjusted to operate. If it is the particular value of i5 at which the relay operates, (17-5) It can be seen from this figure that for a given relay and circuit voltage, the operating time can be controlled by varying Ip, by changing RC, or by changing the balance between R and C while keeping RC constant.

With all these variables to work with, the• principal aim of the design job is to choose the optimum values of the circuit constants to

Circuits

for Timing

387

give the required time with m1mmum variation of time. To meet this objective, of course, circuit elements with low tolerance limits must be used. Principles of Design: Factors Affecting Accuracy, ( 17-1) the secondary current is through the relay is:

From equation

(17-6) Solving this equation for the value of "time" T atwhich the relay should operate, the result, expressed in terms of base-10 logarithms, is (17-7) value of is at which the relay operates. This where it is the particular means that T varies directly with Rs, C, and the log of the ratio of E to it Rs. Except when the operating point is chosen so that T = RsC, there are two values of Rs which will give the required time. Of these, only

n

AMPERE TURNS(N

.p'I ~

l~\

~ -)

C

I

t

\ I '

-

s p

I ..... .".

-

r

I/ II

I

~

! •

Ns is

'l'

"\'

\

--- ..~-\~--1

(Nplpl (Hplp-N

--

----

K..,

;.\--'

l) --

\

\

----

,---· r-•

\~

\

'\'

-

..."

"

,I

0

t

RELAY OPERATES

Fig. 17-5 Curves Illustrating

IN sis CURVES FOR DIFFERENT VALUES

OF RC l

-- -- -r---. ..... __

.1__: r---....

......... 0

,_/

............. V - ...

TIME

the Operating Polnt of the Tlmlng Relay

---

The Design of Switching

388

Circuits

the lower value should be used, since the rate of change of discharge current (and hence the accuracy) is greater for the low R 5 • To obtain best results, the secondary timing winding of the relay should be as effective as possible and should have an exact number of turns of wire. This means crowding as many secondary turns as possible on the relay, consistent with permitting the primary winding to function. The primary winding also should have an exact number of turns. To reduce the variation in the timing obtained by the relay, the lowest possible operate adjustment and the highest possible non-operate adjustment is desirable. This reduces the variation in (Npip - Ni), the ampere-turn level at which the relay operates. Variation of Ni (operate adjustment ampere turns), causes corresponding variation in time from the specified design value. The primary ampere turns, which hold the relay operated after the secondary current has died away, must be several times the operate test ampere turns to provide adequate contact pressure. A reasonable value for Np Ip is five or six times the operate adjustment ampere turns. Since the operating point N 5 it of the relay is equal to Np Ip - Ni, Nl should be as small as possible in order notto raise it too high for flexibility of design. When the variation in circuit elements and the applied voltage is taken into account, equation (17-7) can be modified to represent minimum time and maximum time: E(max) T(mln) = 2.3 Rs(mln) C(min) loglO -----i t(max) Rs(min)

( 17-8)

E(min) R 1 t(min) s(max)

(17-9)

T(max) = 2.3 Rs(max) C(max) log 10 .

In equation ( 17 -9): E(min) where: Ip(mln) = Rp(max) (17-10) In equation ( 17 -8): Np Ip(max) -Ni(N.O.Tst)• . lt(max) = Ns

where:

Ip(max) = R

• Typical examples of the operate (Opr. Tst) and non-operate adjustment values are shown ln the table. of Fig. 17-4.

E(max) ( 17 -11) p(min)

(N.O.Tst)

ampere -turn

Circuits

for Timing

389

Although the effects of variation in the applied voltage tend to cancel out, an increase in E reduces the net operating time through its modification of the value of it . It should be noted that the direct effect of Rs upon T is greater than that of its effect through the log ratio, and therefore Rs(min) is used throughout the equation for T(mtn), and similarly Rs(max) is used in the equation for T(max). With this information in mind, there are several factors affecting the design which can be demonstrated mathematically. Without going through the mathematical steps, the most pertinent factors are as follows: (1) For any particular value i 1 of the secondary current at which the relay operates, the minimum capacity that can be used to give a time delay T is equal to T C = Rs where Rs

(17-12)

(t=2.718}

Therefore, the lowest possible capacity for a given time with a given relay can be determined by calculating the minimum allowable it from equation (17.-10), and from equation (17-12) determining Rs and C. However, if minimum time variation is desired, other factors make a different approach desirable. variation of time, with a given value of RsC, (2) • For minimum Rs should be as low as possible and C as high as possible. (3)• If the value of Rs is fixed accuracy is obtained by setting

as a starting

C = ~s (to give the operating

point,

the greatest

point at T = RsC}

( 17-13)

slnce the charge curve based on this value of C has a greater slope as it passes through time T than the curve for any other value of C. (4)• If the value of C is fixed as a starting curacy is obtained by setting Rs =

}c(to give

the operating

point, the greatest

point at T = 2RsC}

ac-

(17-14)

since the charge curve based on this value of R has a greater slope as it passes through time T than the curve for any other value of R. • The derivation of the relationships In items (2). (3), (4), and (5) ls basedon finding the curve with maximum slope through the operating point. Variations In circuit constants Whenthe slope is maximum will result In least variation in time. These results can be checkedroughly by sketching exponentialcurves for several values of the variable in question on a co-ordinate graph whoseabscissa ls time Insteadoft/RC.

The Design of Switching Circuits

390

(5)• The greater the circuit voltage, the greater is the accuracy that can be attained. However, good accuracy is obtained by using a battery supply of 48 volts. Items (3) and (4) may appear at first glance inconsistent in that for the two conditions of Rs and C fixed, they indicate two different points on the exponential curve as optimum for minimum variation of time. However, this apparent inconsistency merely means that if a value of Rs is chosen, and then the best value of C is calculated for that Rs, from the calculated value of C a new Rs can be determined which will give more accurate results. Successive applications of the rules of items (3) and (4), starting with an arbitrary value of Rs and the desired time T, will result in successively lower values of Rs and higher values of C, to agree with item (2). In designing a circuit to meet certain time limits, there are different methods of approach. Since there are several elements that can fixed in order be varied, something (either Rs or C) must be arbitrarily to start the design. Economic considerations may determine the maximum capacity that can be used, in which case C is fixed and Rs is calculated to give an operating point such that T = 2RsC. U the closest possible time tolerance regardless of amount of capacity is desired, the is smallest Rs (relay winding plus small contact protecting resistance) taken as fixed and the capacity should be calculated to give the operating point at T = RsC. Since the latter procedure usually results in a high value of capacity, and does not give an appreciable improvement in accuracy over T = 2RsC, the customary procedure is to set Rs at its minimum value and choose C to put the operating point in the vicinity of T = 2RsC. This is the method that will be used in the example of designing a circuit, given later in this chapter. It should be noted that if the design is started with a fixed value of C, the resultant Rs must fall within certain limits to be practicable. That is, Rs cannot be less than the secondary winding resistance of the relay, and cannot be so great that the level of it requires primary ampere turns less than five or six times the operate test value. In either of these eventualities, it is necessary to change the assumed initial circuit considerations. The considerations thus far have been concerned only with the theoretical aspects of the problem and have assumed zero inductance, zero armature travel time, and no contact chatter. Inductance and travel time generally have negligible effect when the delay time is high, say • See the footnote on page 389.

Circuits

for Timing

391

above 0.1 second. Below this level, however, these factors may increase the calculated time by a considerable percentage and may have to be taken into account. There is no hard and fast rule for calculating the magnitude of the extra time incurred by these factors, so that if it is desired to keep the minimum delay time below a fixed level, a laboratory test of the theoretical design is in order. Of even greater importance is an apparent change that occurs in the operating current value of the relay. The adjustment procedure provides for a sudden application of current. In a working circuit, however, the current change in the relay is gradual as the operating level is approached. The chief result of this is that the relay does not operate instantaneously as the current passes through the theoretical operating point. In effect, the operating adjustment value of the relay is increased. In practice, this effect may be compensated for by adjusting the design value of primary resistance to give the true time as measured in the laboratory. Still another factor which has to be considered in practice is the change in value of circuit constants with temperature. For instance, if the timing circuit is to be used at frequent intervals, or stands with battery and ground connected, the resistance of the relay windings and circuit resistors will probably change due to temperature rise. Also, a difference in ambient temperature from the rated level will affect circuit values. This can have considerable effect upon timing accuracy unless accounted for in the design. The amount of chatter involved for a given design is also difficult to calculate in advance. Sensitive polar relays are likely to produce some chatter which varies at different operating current values, so that this factor should also be checked with a test model of the circuit. Method of Design. The information given in the preceding paragraphs is adequate to permit the design of a capacitor-timed relay circuit. However, to illustrate the technique involved, a specific circuit will be designed. A frequent situation calling for this type of circuit requires a definite minimum delay time, with some degree of flexibility as far as maximum lime is concerned. The example will therefore design for the minimum time T, starting with a minimum value of Rs (relay resistance plus a nominal rs = 250 ohms to protect the contact which discharges C). The operating point will be set approximately at T = 2R 5 C, although picking commercial values of one-per-cent capacitors may change the operating point slightly. The change in calculations necessary to design for average time instead of minimum time should be obvious from the example.

The Design of Switching Circuits

392

Since, in the general design problem, the object is to determine Rs or C, and Rp, starting with the value of T, equations ( 17 -8) and (17-9) are not the most useful. The most convenient formula is the basic equation ·

lt

_ E r-T/RC jf..,

-

in conjunction with equations (17-10) and (17-11), in which it and Ip are related. Equations (17-8) and (17-9), however, are useful in that they show where maximum and minimum values should be used to calculate maximum and minimum time. In using the exponential form of the equation, the curve of Fig. 17-3 can be used with advantage, although greater accuracy can be achieved by use of a table of exponentials. Problem: To design a circuit giving a minimum delay of 0.060 second. The following are the pre-determined (1) Circuit:

factors:

that of Fig. 17-4.

(2) Relay: that of Fig. 17-4. (3) E = 45 to 50 volts; normal 48 volts. (4) Rs = 2600 ohms + 250 ohms = 2850 ohms (2600 ohms ls the winding resistance and 250 ohms is the external resistance). (5) Operating point approximately The approximate

T

= 2RsC,

value of C (from T = 2R 5 C) is:

c = 2 ~0: 850 = 10.53

mf.

The standard one-per-cent paper capacitors that will be used insolving this problem come in multiples of .540 mf. Therefore the combination of standard capacitors nearest to 10.53 mf should be chosen. Let C = two capacitors at 4.32 mf plus one capacitor at 2.16 mf = 10.80 mf. Now, calculate

the constants to give minimum time = 0.060 second.

0,

Minimum RsC = .99 x 2850 x .99 x l ~~ 8 ~00 = .0302 (both Rs

,

and C assumed

1% low)

T .06 1 99 RsC = .0302 = '

Circuits

for Timing

!:

from

393

curve A, Fig. 17-3, or table of exponentials=

0.136

E(max) _ 50 _ . _ lt(max) - .136 R . - .136 x 2850 99 - .00241 amperes. s(min) X • To obtain lp(max) and Rp(min)

2000 lp(max) = ( 14000

X

.00241) + 1.8

lp(max) = .01777 E(max) 50 Rp(min) = 1 = _01777 = 2810 ohms p(max) Let Rp = 2840 ohms -t1% Resultant

circuit

constants:

rs = 250 ohms :!:1% (Rs = 2850 ohms ±1%) rp = 2640 ohms ±1% (plus 200 ohms relay winding to give Rp = 2840 ohms ±1 %) C = 10.80 mf ±1% To calculate

maximum

time:

Compute it(mln) using equation (17-10). First:

E(mln) 45 lp(min) = Rp(max) = 2840 x 1.0l = .0157 ampere.

. . _ Nplp(min) - Ni(Opr.Tst) Hence. 1 t(min) -

Ns

or: i

t(min)

= (2000 x .0157) - 6 = 001814 ampere 14000 • •

In order to calculate

T(max), equation (17-9) may be used: E(min)

T(max) = 2.3Rs(max) C(max) log or:

10

i

t(min)

R

s(max)

T(max) = 2.3 x 2850 x 1.01 x 10.80 x 10-6 x 1.01 x loglO .001814

45 X 2850

X

1.01

The Design of Switching

394

Circuits

Hence: T(max) = .0676 second. To calculate

average time:

48 = .0169 2840 p(avg)

E(avg) lp(avg)= -R--=

. = (2000 X .0169) - 4 = 002128 lt(avg) 14000 •

T(avg)= 2.3 x 2850

X

10.80

X

10-S loglO _00212

!\

2850 = .0636 second.

It has been noted earlier that the effective operating currentlevel of the relay may be shifted by the nature of the circuit. Also, there may be slight inductive and chatter effects. Therefore, the circuit should be tested in the laboratory, and the value of rp should be adjusted, if necessary, to give the correct time.

41-= _ ____,

7---

B

:I:

:::::c Flg. 17-8 Clrcult Variations

C of Capacltor-Tlmed

Relays

Circuits

for Timing

395

Other Circuits. There are several variations of the basic capacitor-timed relay circuit, some of which are shown in Fig. 17-6. Fig. 17-6B takes advantage of pre-charging the capacitor in order to obtain a higher effective voltage in the secondary winding circuit. The circuits of Figs. 17-6A, 17-6B, and 17-6C are considerably more complicated to calculate than those of Fig. 17-4 because the primary and secondary circuits of the relay are inter-related. The analysis of these circuits can be most easily handled by the application of Thevenin' s theorem to the networks.

17.4

GAS -TUBE

TIMING

CffiCUITS

The capacitor-timed relay circuit described in Section 17.3 requires on the average about 16 mf for every 0.1 second of delay. This value cannot be greatly reduced because of the desirability of operating at the point T = 2R 5 C, and also because of the limits which Rs cannot exceed due to the current requirements of the relay. The result of this is that a delay time above a few tenths of a second requires excessive mounting space for capacitors and is, in general, uneconomical. There is another commonly used capacitor - timing circuit, however, which avoids certain of these limitations, although with some loss of accuracy. This circuit employs a three-element cold-cathode gas tube which operates from the exponentially rising capacitor voltage, rather than from the capacitor charge or discharge current. The gas tube acts as a trigger device which uses for the timing function two of the tube elements, and then operates the delay relay through the third element. In this way the winding of the delay relay is not involved in the timing circuit.

A typical circuit of this type is shown in Fig. 17-7. The timing action starts when the control relay operates to remove the short circuit from the capacitor, apply the high voltage E, and ground the cathode (K) of the gas tube. The tube characteristics are such that the main gap will not break down at the voltage E until the control gap has broken down. The control gap, however, breaks down when the capacitor voltage, applied between the control anode (CA) and the cathode (K), reaches a critical level. For tubes normally used in such application this is about 70 volts. Current flow from anode (A) to cathode begins almost instantaneously with the breakdown of the control gap. The voltage across the capacitor is applied to the control anode while the cathode is grounded, and it rises exponentially with time at a rate determined by Rand C, following curve B of Fig. 17-3. The equation for this curve is (17-3)

The Design of Switching Circuits

396

If the control gap breakdown voltage is represented breakdown is

by et, the time until

E T = 2.3 RC log 10 -E-or - et

1 T = 2.3 RC log 10 l _ et/E

(17-15}

In the circuit of Fig. 17-7, current flow via the anode (A} and the delay relay winding starts immediately after control-gap breakdown, and the delay relay operates. The relay can hold to the gas-tube current or, as shown in the figure, lock to a separate winding and open the gastube circuit. The total delay time is that given by equation ( 17 -15}, plus the relay operating time. Since the delay relay is usually a generalpurpose relay, its action time may be appreciable in short-time circuits, although when the total delay time is of the order of half a second or more, the relay acting time can be neglected. The 100,000-ohm resistance connected to the control anode does not affect the timing, but is necessary to limit the control-gap current at the moment of breakdown. The 1000-ohm resistor, whose function is to limit the current through the contact that discharges C, also has no practical effect on the timing. Since the common cold-cathode tubes require a relatively high operating voltage, this circuit cannot be used where the supply voltage E --4-

t

I

0~

CONTROL

BREAKDOWN VOLTAGE•8t

8c;•E(1-t·t/Rc.) E

t•2.3 RC LOG E•et

Flg. 17-7 Capacltor and Gas -Tube Tlmlng Clrcult

Circuits

for Timing

397

is limited to 45 to 50 volts. However, sources of 125 to 135 volts or more are widely available, and are well-suited to the requirements of this circuit. Factors Affecting Accuracy. Examination of equation (17-15) will show that the time varies directly with Rand C, but decreases as E increases. For a fixed E, time varies directly with et. Therefore, the timing equations can be written: E(min) T(max) = 2.3R(max) C(max) l og 10 ----E(min) -et(max) T(min) = 2.3R(min) C(min) loglO E

E(max) (max) -et(min)

(17-16)

(17-17)

In addition, time increases with the operating time of the delay relay. This factor is negligible for long time intervals. All circuit constants should be held to as close limits as possible for accurate timing. This implies that capacitors and resistors be as accurate as is economical. Variation of the 130-volt supply by t5 volts is equivalent to ±3.8 per cent. The control-gap breakdown voltage of the tube depends upon the type. For accurate timing, a tube type with the lowest possible variation in control-gap breakdown voltage should be used. It should be kept .in mind that the characteristics of a cold-cathode gas tube tend to change with time. The slope of the ec curve through the operating point also affects the accuracy of the timing. Theoretically, the optimum slope for a given supply voltage occurs when the operating point is set such that T = RC. However, the value of the nominal firing voltage (70 volts), when used in conjunction with a supply voltage of 130 volts, serves to give a nat-:ural operating point at about T = .75 RC, since the capacitor is 54 per cent charged at 70 volts. The minor difference in accuracy between this point and T = RC usually makes it unnecessary to attempt to set the operating point at T = RC by means of bias voltages. The slope of the ec curve through the operating point is also directly affected by the charging voltage, increasing as the voltage is set at higher levels. For this reason an improvement in accuracy can be obtained by using either a higher positive potential or a negative voltage supply to pre-charge the timing capacitor to oppose the +130-volt charging voltage. Care must be taken to insure that the effective supply voltage does not exceed the main-gap working voltage. Theoretically, for a fixed value of RC, the ratio between R and C makes no difference. However, in the interests of economy it is desirable to hold the value of C as low as possible and increase R, subject to an upper limit as discussed on the following page.

The Design of Switching

398

CIRCUI.T WITH LEAKAGE RESISTANCE

Circuits

RL

E'---1\

EQUIVALENT CIRCUIT WHERE E1 =

~E R+RL

1_ RRL R - R+RL

Fig. 17 -8 Method o( Accounting for Leakage Resistance

A factor that affects the accuracy is the relationship of the value of R to the possible leakage resistance which may appear shunted across C. Leakage resistance may drop to as little as ten megohms under service conditions, largely because of the effect of dirt between capacitor terminals when the humidity is high. Some leakage may also be contributed by the wiring and the spring pile-up of the control relay. The effect of leakage resistance, illustrated in Fig. 17-8, ls to deIn gencrease the effective charging voltage and charging resistance. eral, this results in an increase in time. One method of diminishing the effect of leakage resistance is to hold the charging resistance R to a maximum of about one megohm. At this value aleakage of ten megohms increases circuit time by about six per cent. Another method is to arrange the circuit in such a manner that time is independent of leakage resistance down to a value approximately the same as the charging resistance itself. This is feasible because, as was already noted, leakage reduces both the effective charging voltage and the charging resistance, and equation (17 -15) If, before the shows that these effects are partially self -compensating. timing cycle starts, the capacitor is set to a potential opposing the charging voltage, a point can be found on the charging curve where the values of ec and T are independent of leakage resistance, RL, down to a low value of the latter. The effectiveness of this method depends upon

Circuits

for Timing

399

making sure that the leakage is to a known potential such as ground; otherwise the correct operating point is indeterminate. A circuit for accomplishing this is shown in Fig. 17-9, together with a sketch (not to scale) which illustrates the principle involved. The bias value in this circuit is applicable when the available supply volt.: ages are +130 volts and -48 volts, and requires some modification for other voltages. The timing equation is derived from the exponential form given in (17-3) of the footnotes on page 384: T = 2.3 m RC log

10

mE + ex E - et

(17-18)

m

where m is a factor RL / (R + Rd due to leakage. The circuit is arranged so that leakage is to ground by grounding one side of C and placing ground on a make-contact of the relay spring pile-up to which the control anode of the tube is connected. The capacitor is pre-charged to-48 volts through a low resistance (negligible compared to Rd. Under these conditions, the time at ec = 30 volts is unaffected by the value of RL (the factor m), as long as RL is greater than about 2R. Therefore, the cathode bias on the tube must be set at -40 volts so that the tube fires at ec volts. E•+l30V

__clI-- -i -loo---

~

~n_---R--

RL R+RL•l30V

t-----7~=========-..

WITH RL

30V

·48V

Flg. 17-9 Clrcult to Ellmlnate Effect of Leakage Resistance upon Tlmlng Accuracy

400

The Design of Switching

Circuits

The curves sketched in Fig. 17-9 indicate the action of the circuit. During the early part of the charging period, the effect of RL is to increase the rate of charging; during the latter part, the effect is to decrease the rate. At ec = 30 volts, the two effects cancel. Principles of Design. The previous paragraph discussed the essential principles involved in the design of the capacitor gas - tube timing circuit. To design a circuit for an average delay time T, the tube would be chosen first and the average value of the firing voltage et, obtained. Calculated from this and the circuit voltage E, the ratio et /E can be used to evaluate T /RC, using curve B of Fig. 17 -3 or a table of exponentials. From T /RC can be calculated the approximate value of C corresponding to the known v_alue of T and R = one megohm. The value of C nearest to this figure which is obtainable in standard one-per-cent capacitors can then be obtained. The actual resistance of R can then be recalculated from the value of T /RC. Using these values of R and C, the minimum and maximum times can be obtained using equations (17-16) and (17-17). In calculating maximum time, it is necessary to take account of leakage resistance unless the self-compensating circuit is used. The life of the gas tube varies inversely with the anode current, and the current values that should not be exceeded are listed under the tube characteristics. Therefore, it is necessary to insert in series with the main gap sufficient resistance to limit the anode current to a reasonable value (usually peak .030 ampere, average .010 ampere). The available voltage in the external main-gap circuit is the difference between the supply voltage and the main-gap sustaining voltage, the latter being generally of the order of 70 to 80 volts. In general, the relay resistance should be as high as possible, and in some cases additional resistance should be inserted. It is usually advisable to open the tube circuit when the relay operates in order to lengthen tube life. When choosing the relay it is necessary, of course, to make sure that its operate current requirement falls within the worst circuU margins. (minimum available relay voltage equals minimum supply voltage minus maximum sustaining voltage). The accuracy of the timing circuit is subject to the variations encountered in voltage supply, tube constants, and resistance-capacitance values. With a voltage supply of 125 to 130 volts, accuracy of about ± 20 per cent can be achieved. Variable delay can easily be obtained by making R or C variable in the circuit. Roughly speaking, the circuit requires about 1.33 mf per second when R is set at one megohm. Since the gas tube is a high-impedance device, the leads connected to the tube terminals are subject to induced voltages which may be high enough to fire the tube falsely unless precautions are taken.

Circuits for Timing Two types of cable transients are recognized, phenomenon, the other of low frequency.

401

one a high - frequency

The high-frequency pick-up generally occurs on the long leads carrying battery and ground to the circuit. It occurs when the tube is in close proximity to a lead connected to an unprotected relay which releases. These transients can be protected against by a .001 mf mica condenser connected directly to the tube socket terminals. The induced low-frequency transient occurs on leads connecting the tube and a relay winding. Trouble can occur with leads as short as two feet, so that it is important to keep relay and tube very close together. If a tube that is not fired is left with an unterminated lead connected to the tube anode or cathode, charges may accumulate on the lead that will result in false firing of the tube. A ten-megohm resistor should be connected permanently to such a lead to drain off the charges.

A so-called enabling transient high enough to fire the tube falsely may occur when the high voltage is connected across the main gap of the tube at the start of the timing period. This depends upon the circuit configuration and constants, and the design should be scrutinized carefully to insure that no trouble arises from this cause. A voltage induced in the relay winding by magnetic interference may also fire the tube prematurely. Appropriate precautions against magnetic interference must be taken, therefore, and in addition, at least 40 volts margin should be left between supply voltage and main-gap breakdown. Four circuit arrangements designed for particular types of application are shown in Fig. 17-10, with notes as to the conditions under which they can fail. The circuit of Fig. 17-l0A, which utilizes an extra relay to check the capacitor discharge path, cannot fail by false or premature firing because of dirty contacts in the control path. The circuit of Fig. 17-10B is arranged so that dirty contacts in the control path cannot prevent the tube from firing. It is suitable for use for certain trouble -detecting purposes where failure to give a trouble indication cannot be tolerated. The tube is extinguished after the timing period by the shunting action of the relay locking contact. The voltage surge developed at the release of the relay necessitates use of the 2500 ohms - 2 mf network across the tube to prevent false firing across the main gap. The circuit of Fig. 17-l0C is subject to failure by either false firing or nonfiring because of a dirty contact in the control path. However, it is the simplest and most economical of the several circuits and is recommended for general timing purposes. Fig. 17-100 is a modification of Fig. 17-lOC for use where a less expensive tube with high

The Design of Switching Circuits

402

main-gap working voltage can be used instead low-voltage tube.

of a closer

tolerance

In specific applications of any of the various timing circuits, care must be taken that sufficient time is allowed between uses to permit the circuit to return completely to normal. If the tube has not completely de-ionized, the capacitor discharged, and the relay de-energized, the next timing interval will be in error. The restoration time required +130V

__c(~---1

R

fJ

NORMALLY -r C CLOSED ~

_r_----o---

DIRTY CONTACT CANNOT CAUSE FAILURE BY PREVENTING FIRING B

DIRTY CONTACT CANNOT CAUSE FAILURE BY FALSE FIRING. (CHECK RELAY GUARD)

A +130V

~~---t R 0.001.uF

O.OOlµF C CONTROL

D 1000.n.

-sovt DIRTY CONTACT CAN CAUSE FAILURE BY NONFIRING OR FALSE FIRING

DIRTY CONTACT CAN CAUSE FAILURE BY NONFIRING OR FALSE FIRING

C

Fig. 17-10 Alternative

D Gas-Tube Timing

Circuits

Circuits

for Timing

403

depends upon the particular relation to the requirements

circuit and should always be determined of the application.

in

Design Example Problem: to design a circuit onds with deviation within about 17-IOC is satisfactory.

to give an average delay of ten secof Fig.

± 25 per cent. The circuit

Voltage: E = 125 to 135 volts Control-gap breakdown: 69 to 74 volts Leakage resistance: 10 megohms minimum. The value of R should be approximately one megohm. The tube breaks down at 72 volts average, and the average supply voltage is 130 volts. Therefore:

Eet

72

= 130 = .554

T From curve B of Fig. 17-3 the value of T /RC is seen to be: RC = .80 Substituting R = one megohm and T = 10, the value of C is 12.5 mf. The value of three 4.32 mf ± 1 per cent capacitors is 12.96 mf, and this capacitance will be used. Substituting this value in the equation for T /RC gives a value for R = .965 megohms. It will be assumed that a -t 5 per cent resistor will be employed in the circuit. If the series relay is 2500 ohms, a .96-megohm resistor can be used. The value of T(mln) can now be calculated using equation (17-17). An alternative method would be to use equation (17-3) and curve B of Fig. 17-3, substituting minimum and maximum values as indicated in equation (17-17). Substituting appropriate values in equation (17-17) gives: 135 T(min) = 2.3 X .9625 X .95 X 12.96 X .99 X log10 135 _69 or:

T(mln) = 8.36 second

The value of T(max) is obtained from equation (17 -16). However, first must be calculated the equivalent E and R which appear due to leakage. Employing the equations of Fig. 17 -8:

E' (min) = R

E' (min)

=

(max)+

R

L

x E(mln)

10 _ (_9625 x l.0 5) + 10 x 125 - 113.5 volts

~max)RL .9625 X 1.05 X 10 R'(max) = ~max) + RL = (.9625 x 1.05) + 10 = .918 megohm.

The Design of Switching

404 Substituting

Circuits

in equation ( 17 -16}:

T(max) = 2.3

X

.918

X

12.96

X

113.5 1.01 log 113 _5 _74

Tcmax) = 12.7 seconds. The relay operating time in this case would be negligible. The circuit constants

required

are:

R = .96 megohm ±5 per cent C = 12.96 mf ± 1 per cent and they give a time range of 8.36 to 12.7 seconds.

PROBLEMS FOR CHAPTER 17 17-1 Design a capacitor-timed relay circuit to give an average interval or 0.025 second with as small a variation as is practical. Use a relay or the type shown in Fig. 17-4. Assume that capacitors or .±1% accuracy are available in values which are multiples or 0.540 mr. A battery supply normally held at -48 volts, but which may range from -45 to -50 volts, is to be employed. Assume that the average ampereturn value at which the relay operates is 4. Determine 17-2

theoretical

upper and lower timing

limits

for the circuit.

An Interrupter contact (INT) operates as shown tn the diagram below. Using this contact as a timing mechanism, design a relay circuit to ground a lead TM not sooner than 60 seconds nor later than 75 seconds after a start lead ST ls grounded. Lead TM shall remain grounded thereafter until ground ls removed from lead ST. The removal of ground at any time from lead ST shall restore the circuit to normal. The transfer time or the interrupter contact ts less time of a relay. (This can be done with five relays.)

than the operate

or release

(INT)

B

F

TM

RELAY

B

F

CIRCUIT ST

_j_

-

Chapter 18 CIRClnTS FOR PULSE GENERATION As mentioned in the preceding chapter, circuits for pulse generation are closely allied to timing circuits. This similarity follows from the function of a pulsing circuit: the generation of recurring timed intervals in the form of a train of pulses. A pulse generator can be considered as a recycling timing circuit which alternately produces ·measured pulse intervals and inter-pulse intervals. The need for recurring pulses arises at many points in the design of switching systems. For example, information is often passed from circuit unit to circuit unit, or from location to location, by means of trains of pulses. These pulses may consist of opens or closures of a signaling loop, or of ground or other potential condition on a single lead. Pulse trains are also utilized to time critical operations and to control time sequences. In some applications the pulse cycle must be accurately timed; in others, considerable timing variation may be allowed. The required repetition rate may be measured in pulses per minute, pulses per second, pulses per millisecond, or, in some cases, in pulses per microsecond. Pulses can be generated by a variety of methods: by mechanical devices, by motor-driven interrupters, or by relay and electronic circuits. The form of circuit and the circuit components to be used depend upon the speed of pulsing and the accuracy of pulse and inter-pulse intervals necessitated by the circuit requirements. In this chapter, pulse generators based on relays and on tubes are considered, together with means for detecting pulse information. 18.1 SIMPLE RELAY PULSING CIRClnTS The self-interrupting or "buzzing" relay circuit, so undesirable in most circuit applications, lends itself to pulse generation in those cases in which considerable variation in pulse length and repetition rate is permissible. Its operation is based primarily upon the principle of closing ground to a relay through the break-contact of another relay which is itself operated, directly or indirectly, by the first relay. The relays will buzz or pump as long as the initiating ground-closure is maintained.

- 405 -

The Design of Switching Circuits

406

~ dJ

IA)

...r:-

-

I

__ _,

--l1I

~

-=-

!]_ -=-

----~>-n---·

-!- 1

11+--:i_

}

(B)

.

PULSING LEADS TO EXTERNAL CIRCUIT

RELAY IA) RELAY (Bl

A } -------------

PULSING LEADS TO EXTERNAL CIRCUIT

IA)

'+--+----.J\Af\,-.-411r-i

-!=t •

♦

= lir-::i._

(Bl

RELAY IA) RELAY IB)

B

er----} (Al

PULSING LEADS TO EXTERNAL CIRCUIT

RELAY (A) RELAY (Bl RELAY IC)

C Flg. 18-1 Simple Relay Pulse Generators

Circuits for Pulse Generation

407

Three examples of this type of circuit are shown in Fig. 18-1. There are many variations of these circuit forms, and the number of relays involved in each can be extended to obtain the desired intervals between pulses. The intervals obtained are, of course, directly dependent upon the operating and release times of the relays, and are accurate within the variations to which relay acting times are subject.

18.2 CAPACITOR-TIMED

RELAY PULSE GENERATORS

The capacitor-timed operation of a suitable relay as a means for prO

,a,,o

C

C

L

,a,,o

(8 1 10

(8410

e,,. ~--

••••□ I A

~--e,,. Ne1•

-F- --

Bo

a,

Bz

8:,

84

84

Bo

81

Bz

8:,

A:,

Az

••

Ao

A4

B

Flg. 21-t

Contact Network for All Qulnary Addltlons Glvlng a Sum or Three

Circuits

467

for Calculating

Bo

B,

Bz

B3

B◄

B◄

Bo

B,

Bz

B3

Ao

A4

A3

Az

A,

A,

Ao

A4

A3

Az

Az

A,

Ao

A4

A3

A3

A2

A,

Ao

A4

A4

A3

Az

A,

Ao

Co

- c, -

Cz

-

Fig. 21-5 Complete Contact Network for Quinary Addition

--

Ao

A,

A2

A3

Bo

Bo

Bo

Bo

e,

e,

e,

Bz

82

A4

-

L5

B:,

--

84

83

82

e,

Bo

A3

A4

e,

Bo

A Ao

A1

A;

84

Bs

82

7

5N

e,

84

82

82

83

83

83

84

84

84

-

B Flg. 21-6 Networks for "Less than 5" or "5 or More"

468

The Design of Switching

Circuits

table, the grounding of L5 and 5M depends only on the values of AQ and BQ, without regard for the conditions of leads NCIN and C1 1N. The L5 and 5M contact networks corresponding to these combinations, then, need pass only through contacts on the (A-) and (B-) relays. On the other hand, contact networks corresponding to sums lying on the diagonal require control by the NCIN and ClIN leads. For example, with (A 3 ) and (B 1) relays operated, L5 is grounded when NCIN is grounded, but 5M is grounded when ground is supplied to C 11N. Suitable relay contact networks to control the internal-carry leads may be drawn by inspection of the table of Fig. 21-3. These contact networks are shown in Figs. 21-6A and 21-6B.

e, 00

5

_........,t-----,t------.1

00

-Aa UPPER LEFT CORNERS

=

LOWER RIGHT CORNERS ""

+ = 5

00+

5+

NO+

•

L5

GROUNDED

5M GROUNDED ClouT GROUNDED NCour GROUNDED

C1 UN CENTRAL BLOCKS)

Fig. 21-7 Table for Binary Addition

The Binary-Digit Adder. The function of the binary digit adder is to determine the sum of the binary digits, AB + BB, taking into account the carry information from the quinary adder, and also to ground one of two leads, NCouT and C louT. The digits are stored in the A and B registers by the operation of relay (A 00 ) or (A 5) and relay (B 00 ) or (B 5 ), the subscripts 00 and 5 corresponding to the binary digits. The binary sum, similarly, appears on either l.·ad c 00 or lead C 5 • The carry-in digit from the quinary adder is either O or 5 depending upon the condition of leads L5 and 5M, respectively. A table for binary addition, similar to that used for quinary addition, is first written down as shown on Fig. 21-7. The table is then used B5

8 00

B5

- Coo

85

_r---

NCouT

8 00

Aoo

5M

....r:.---

-

Boo

A5

Aoo

L5

A5

Aoo

ADDING NETWORK

*

Clour 85

Boo

~ As

5M

__c----

Aoo 00

As

As

CARRY-OUT

B

A Fig. 21-8

Networks for Binary Addition and Carry-Out

NETWORK

Circuits

for Calculating

469

in the same manner as the earlier circuit sketched on Fig. 21-SA.

table to develop the binary addition

Carry-out Network for Binary Adder. The plus signs in the blocks of Fig. 21-7 indicate the combinations which must ground the C 10 UT lead, while the blocks without plus signs represent the combinationsfor grounding NCouT· Two of the digital combinations, that for A00 + B 00 and for A 5 + B 5 , ground NCouT and ClouT respectively without regard to the value of the input carry leads, and therefore can be developed with local grounds. The remaining digital combinations must be taken in conjunction with the L5 and 5M carry inputs to produce the output carry signals. The complete carry-out network is shown on Fig. 21-SB. Complete Adder Unit. To arrive at the complete contact network for a single stage of biquinary addition, it is merely necessary to combine the four component networks derived in the preceding paragraphs. This is accomplished as shown in Fig. 21-9. Recalling the over-all circuit action originally specified, the two numbers to be added are C1N ~-=- 84 N

....c--'~C

Bo B1 Bi 83 84

83 B2 B, Bo 84

Bi B2 81 Bo

Bo Bo Bo Bo

~I

--=83"'

84

f,-

Bo B1 82 83

A1 Ac A4 Ai Ai

83

83 83

A2 A1 Ao A4 A3

84

84

84

Co

Ao A.i A3 A2 A1

82

8,

B, B1 81 B2 B2

84

-:.:- Al A2 A1 Ao

A4

-

A4 A3 A2 A1 Ao

Coo

Boo 8:) 85

Boo Boo 85

85 Boo

A~ 5M Ao Ao

A,

Flg. 21-9

Az

A3

A4

A1 A2 A3 A4

L5

r1t,

A5 ~()(

1'ooA1

Aoo Aa Aoo A1

Complete Adder Network Unlt for Single Blqulnary Palr Summation

ClouT

The Design of Switching Circuits

470

placed in biquinary registers, a single register unit for each decimal digit. The addition is performed by summing each corresponding pair of decimal digits. The contact network of Fig. 21-9, controlled by the register relays, carries out the summation of two biquinary pairs. A necessary supplement to this network is, of course, a contact arrangement to check that one and only one of the binary relays and one and only one of the quinary relays in each register unit are operated. An adding circuit may be arranged to operate on numbers of any desired number of digits by connecting these single decimal-digit-adder SEQUENCE CIRCUIT

r--

r-

I..---I

,- - --1

ADDER-7

I

REGISTER A

1

I

....... - CONNECTOR - _________

._ ___

I + I PRIMING I REGISTER 8 ..-1--------4---t-t-___ I '--~----' I = = I I REGISTER C I + I I REGISTER D ~-----4 I ___ __, I L ____ _J

..,_ __

..... _,. ______

_,

-I-.----+--'

I I I

L.... - - - - - - - - ·- - J B Fig. 21-13 Simple Multiplication

6 + 6 + 6 = 18. Although a circuit based on this latter method might seem to be unduly slow in obtaining a final solution, it is, on the average, fully as rapid as a circuit based on the first method. Moreover, the second method may be adapted more conveniently to modifications of the basic adder circuits. Multiplication Employing a Relay Multiplication Table. To provide on which to devise a system of mechanized multiplication, consider the procedure followed in solving a simple problem. For example, we shall multiply the multiplicand 847 by the multiplier 3. U our memory is poor, we first multiply 3 x 7 by consulting a table, and find the product, 21, which we write down. We then find 3 x 4 :a: 12, which product we write, shifted one place to the left, below the first product, 21. We determine the product of 3 x 8 from the table, and write this product, 24, again shifted to the left, below the earlier products. The sum of all products is the solution originally desired. This sequence of actions is indicated by Fig. 21-13A. a framework

This plan may be followed almost exactly by a relay circuit, as shown by the block diagram of Fig. 21-13B, assuming that a relay multiplication table is realizable. This multiplication table may operate as a translator which functions . according to the rule X • Y = Z, where X and Y are decimal digits and z is a number composed of a tens and a units decimal digit. '

476

The Design of Switching

Circuits

847

X3 4 21 2 I 2 2541

----

REGISTER REGISTER REGISTER

A B C

A

----7

r-----

3

ADDER

I 4 MULTIPLICATION

REGISTER A i-- 2-+--

TABLE

8

I

+ 2 3 2

I I

""'

REGISTER C

I I I I I

MULTIPLICATION

7

TABLE 3

I

.J

L ________

B Flg. 21-14

A More Practical

Means for Slmple Multlpllcation

Employing this relay multiplication table, then, the product of 3 x 7 is obtained, and is placed in the A register of the adder. The product of 3 x 4, shifted one place to the left,• is placed in the B register, and the product of 3 x 8, doubly shifted, is set into the D register. The summation of the contents of the A and B registers appears in is register C. As a final circuit action, the number in the D register added to the number now in the C register; the sum obtained in register B is the complete product of 847 x 3. Althoug!l the plan just discussed is a workable scheme, it is somewhat slow·, requiring. three sequential operations of the relay multiplication table and two steps of addition for the problem treated. A more rapid system can be devised at the cost of additional relay multiplication tables which operate -simultaneously. A block of this second system appears in Fig. 21-14B. Here two adder registers are provided; in one is placed a number composed of the units digits of the outputs of • When any number ls shifted be filled with zeros.

to the left, the vacated digital positions

to the rlght must

Circuits

for Calculating

477

the multiplication tables, and in the second is placed a number composed of the output tens digits. The corresponding mathematical procedure is shown in Fig. 21-14A. Note that the digits placed in the B register correspond to "carry" digits. As stated above, the relay multiplication table circuit appearing as a block in Fig. 21-13 and 21-14 is a translator. Although a circuit could be designed to multiply biquinary pairs directly, a less complicated arrangement results from the design of a relay table on a decimal basis. If the latter table is employed, the two decimal digits at the output of the table are translated to pairs of biquinary digits which, in turn, are placed in the conventional adder registers.*

The design of the multiplication translator is straightforward. The final circuit obtained may be of the form shown in Fig. 21-15, in which the input multiplicand digit is introduced by operation of one of the (R-) relays, and the (M-) relay operated corresponds to the multiplier digit. The two digits of the product are transferred to the adder registers, the units digit to the A register and the tens digit to the B register, as indicated by the diagram of Fig. 21-14. The (M-) relays may be of a multicontact type so that several multiplication tables can be constructed on a single set, each table having its own group of (R-) relays. Thus, in the example 847 x 3, the multiplier relay (M3) and the multiplicand relays (RS), (R4), and (R7), in tables 1, 2, and 3, respectively, are operated, the (M3) relay controlling contacts in all three tables. The block diagram of Fig. 21-14 may be expanded into an arrangement which can handle multipliers of any desired number of digits, as shown in Fig. 21-16A. This circuit obtains partial products by multiplying the multiplicand by each digit in the multiplier (as in the preceding paragraphs), and then adding these partial products to arrive at the final product. It is, of course, necessary to shift the partial products depending on the position of the corresponding multiplier digit. As an illustration of the operation of this circuit, the steps involved are shown graphically in Fig. 21-16B, where 435is multiplied by 279. The first operation is that of obtaining the product of 435 x 2 = 0870, accomplished exactly as in the basic block diagram of Fig. 21-14. This partial product, 0870, appears in the C register. Then, the partial product 43 5 x 7 = 3045 is similarly obtained in the D register. The next step directly transfers the partial product, 0870, from the C register to the A register, and transfers the partial product, 3045, from the D register to the B register, shifting the former product one place to the left. The sum of these products, 11745, appears in the C register. The last partial product, 435 x 9 = 3915, is set into the D register. The • An alternative

might involve adder registers arranged (or decimal addition.

The Design of Switching Circuits

478

(M 0)

(M2)

(Mil

...\ ~

(M3l

I @>'- [email protected] @J- l ~

®'

µ

Ill

i:ii .... ct:

.... C ct: C

C GI

0

~

---@ -@

--e -@)

- ,-!

-µ

IM9l

9'

.if rar,, @'

-,.l . - ,-A

9

.ll

-µ

:e l®i1@1 @J ®ij

- f-A ,.....w ~ I:

-'...!

®'

I

8'

(~6J-'µ@Yi

e

-

-µ - µ

,.....µ .,..+l

-

I

!-I ,95\.J .j

>,},

@l (§)-1 @J!~!e1 @,J ! - µ - µ - µ .,....µ - p, - µ -! =(A+

network must now be designed and integrated with From Table 21-3, it is seen that the circuit equaB + C'm)(A + B' + CIN)(A' + B + CIN)(A + B + CIN)

This reduces to: f(CouT> = (A + B)(C IN + AB) or

(B + CIN)(A + BCIN)

or

(A+ CIN)(B + ACIN)

or these

three equivalent circuits, the first is closest in form to the summing network f(C). The similarity can be improved by an additional manipulation which will permit re-use of gates in Fig. 21-23. The final form is: f(CouT)

= (A + B) [ CIN + (A' + B')']

This can be combined with the summing

network

to give the complete

492

The Design of Switching

Circuits

C

B-'---C.N---------t-t--+------

Fig. 21-24 Complete Electronic

circuit of Fig. 21-24. This circuit can be istor "and" gates and three pentode "or" signal inverters which may be transformers circumstances. It will be recognized that but one of a variety of equivalent circuits, oped by algebraic manipulations similar to

Binary Adder

constructed from five vargates, together with three or triodes depending upon the circuit of Fig. 21-24 is any of which can be develthose used in this section.

In high-speed electronic computers, binary numbers are often A--BINARY ---•C represented in serial form by a see-----i ADDER quence of pulses. The time interval Cour for each digit is accurately controlled by a series of "clock" pulses ....__..,. DELAYi---"""' generated by a master oscillator. If the value of a digit is 1, a pulse is Fig. 21-25 Binary Adder for Serial Operation allowed to pass into a lead carrying the number of which this digit is a part; if the digit is 0, the pulse is suppressed. The least significant digit is transmitted first in time sequence. A single binary adder unit, of the type already designed, may be used in conjunction with a timedelay unit to add binary numbers of any number of digits in this serial form. The arrangement is indicated in Fig. 21-25. The carry-out signal is passed through a delay network or other device which causes it to be delayed a time equal to the interval between successive clock ·pulses. Thus the carry from one digit arrives at the input at the same time as the signals representing the following digit of the numbers being added. The sum of the two numbers applied to the A and B inputs then appears as a sequence of pulses on the sum output lead.

Circuits 21.8

for Calculating

APPENDIX

493

I: N2 AS A SUM OF FACTORS

By a simple manipulation

as: n2 = (n - 1)2 + 2n - 1

number n can be written and, therefore,

it may be shown that the square of the

n2 - (n - 1)2 = 2n - 1

Stated verbally, this latter expression indicates that the square of an integer n minus the square of the next lower integer• equals the difference (2n - 1). Thus the square of n may be expressed by a summation of all these differences taken from 12 to n2: n

n2 =

I:

(2m - 1)

1,2,J, ..

n

But

I:

(2m - 1)

is merely

the sum of all odd integers

from 1 to

1,2,J, ..

(2n - 1), and therefore

n2 = 1 + 3 + 5 + 7 + ...

+ (2n - 1).

a

Note that the number of terms in

I:

(2m - 1)

is equal ton.

1,1,3, ..

21.9

APPENDIX Il: DETERMINATION OF SQUARE ROOTS BY REPEATED SUBTRACTIONt

Assume that the desired square root is a three-digit number, 102A + l0B + C. The square of this number is

decimal

104A2 + 2·103AB + 102B2 + 2·102AC + 2·10BC + c2. The first sequence of subtractive steps involves the subtraction of 104 . 1, 104 • 3, 104 • 5 ... in order until the smallest positive difference is obtained. Carrying out this sequence: 104A2 + 2·103AB + 102B2 + 2·102AC + 2·10BC + C2 - 104 • 1 - 10 4 • 3 - 104 • 5

. . . .. . . .. . . - 104 • (2A - 1) First

Remainder

= + 2·103AB + 102B2 + 2·102AC + 2·10BC + C 2

• The expression ls not limited integral values of n.

t Discussion

to integers,

but the method discussed here involves only

of this method ls included in Section 21.6 o( this chapter.

494

The Design of Switching

Circuits

This is the smallest positive difference, since the next subtrahend in order would be 104(2A + 1); and 104(2A + 1), as may be shown, is larger than the remainder above. The digit A is now known, since it is equal to the number of subtractive steps in the sequence. The last-used subtrahend, 10 4 (2A - 1), is modified by adding 4 10 • 1 and dividing by 10 (the equivalent of a shift of one place to the right). This constant factor, plus 102 • 1, 102 • 3, 102 • 5, .... taken in order, is subtracted from the first remainder: 2·103AB + 102B2 + 2·10 2AC + 2·10BC + C2 - (2•103A + 102·1] - [2·103A + 102·3] = - ((2·103A)B

+ 102B2]

- [2·103A + 102(2B - 1)] Second Remainder

= + 2·102AC + 2·10BC + c2

The number of steps in the second sequence is B, and the second remainder is the smallest positive difference, since the next subtractive factor in order would be [ 2 •10 3A + 102 (2B + 1)] , a factor which may be shown to be larger than (2•102AC + 2·10BC + C2). The final group of subtrahends is obtained by adding 102 •1 to the last-used subtrahend in the second sequence, [2·103A + 102 (2B -1)] and dividing by 10. This constant factor plus 1, 3, 5, ... taken in order, is subtracted from the second remainder: 2·102AC + 2·10BC + C2 - (2·102A

+ 2·10B + 1]

- [2·102A

+ 2·10B + 3]

- [2·102A+

- [2·102A Third

2·10B+

5]

= - [(2·102A)C

+ (2·10B)C + C2]

+ 2·10B + (2C - 1)]

Remainder

Since the number

=0 of subtractive

steps in this final sequence is C and the

Circuits

for Calculating

remainder termined.

is 0, the complete

495

root, 1Q2A + 10B + C, has now been de-

By extension it is apparent that the above process holds for numbers with any number of digits.

PROBLEMS FOR CHAPTER 21 21-1

An adding circuit is to be designed to determine the sum of two five-digit binary numbers, A and B, as shown In block form below. Each of these numbers Is represented by ground or absence of ground on five input leads, ground on a lead representing the binary digit "l" and absence of ground representing the binary digit "0". The sum, C, of A + B is to be Indicated in binary form on a group of six output leads in the same fashion. (This can be done with ten relays.)

~

OUTPUT NUMBER C

21-2

Design a circuit to meet requirements similar to those of Problem 21-1, except that the output number C should be the dlf!erence of numbers A and B rather than their sum. It may be assumed that A is normally greater than or equal to B. However, If B should be greater than A, an alarm lead should be grounded. In devising

21-3 Design

a circuit

plan, do not use the method of complementary

addition.

a check network !or a blqulnary register to Indicate that one and only one relay in each of the bl- and qui- parts of the register Is operated. Use only makecontacts !or this circuit, and distribute them so that all qui- relays have the same number of contacts. This circuit was discussed on page 464 of this chapter. (This can be done with 24 make-contacts.)

Chapter 22 GENERAL

PRINCIPLES

OF MULTIFUNCTIONAL

CIRCUIT

DESIGN

A multifunctional switching circuit, as the name implies, comprises any circuit entity that performs more than one circuit function. This broad definition covers almost all circuits that are designed for practical application, since in general the single-function circuits discussed in the preceding chapters are too restricted in nature to be used except in combination. In size and complexity, multifunctional circuits cover the range from large, immensely complicated organisms, such as telephone switching systems or computing systems, down to relatively simple arrangements for automatic control of a machine. However, no matter at what point in the scale a multifunctional circuit falls, it is clear that design techniques enter a new phase over and above what has already been discussed. The question that immediately arises is, what design techniques can be formulated that will be applicable to multifunctional circuits in general. In view of the diversity of functions and applications of switching circuits as a class, there is undoubtedly no specific design technique that is valid for all types of circuit. The nature of the subject is such that almost every design project is unique to some degree and requires creative effort in its solution. The natural result of this is that what is a straightforward and useful method in one case is completely unproductive in another. However, a valuable and universally applicable approach to the design of multifunctional circuits can be enunciated and, if followed intelligently, will yield highly useful results. This approach can best be understood by first analyzing critically the general nature of well-designed circuits, where the term "well-designed" implies an orderly, logical, and straightforward solution. Thus, any switching system can be analyzed circuitwise in terms of interrelated groups of functional blocks, where each block represents a single or closely associated set of single-function switching circuits. There may be intermediate breakdown levels between the system and the circuit blocks, the levels depending, to a great extent, upon the nature and size of the system. A large system, for equipment or utilization reasons, may consist of several major units, each of which is a complex multifunctional circuit. An analysis of a system of this type would show not only the relationship of the major units with each other

- 496 -

General

Principles

of Multifunctional

Circuit

Design

497

and their environment, but also the functional block co-ordination within each unit. A small system, on the other hand, might comprise no more than a few functional blocks related to each other and to the environment. Viewed in this light, the apparatus units such as relays, taken individually, are very subsidiary components of the system. The primary unit of importance is the single-function circuit block. The correct approach to multifunctional circuit design is to some extent the inverse of the analytical operation. From a statement of circuit requirements, a functional plan is developed in terms of known or conceptually evident circuit blocks, representing simple circuits similar to the single-function circuits discussed in the preceding chapters. As the design proceeds, the functional blocks are coordinated and integrated to the point where a comprehensive block diagram of the proposed circuit exists. Only at this late stage, with certain reservations, is the detailed design in terms of apparatus components started. It is this method of approach that will be discussed in the present chapter.

It is beyond the scope of this volume to pursue the study of multifunctional circuit or system design exhaustively or in all its ramifications. For example, the subjects of apparatus design and equipment design, which are intimately related to circuit design, will not be touched upon at all. The many practical aspects which enter into commercial circuit design will only be hinted at. The heart of the matter, from the circuit designer's point of view, will necessarily be treated in very broad and general terms in this chapter. However, in order to illustrate the method concretely, a specific example of the design of a typical multifunctional switching circuit will be presented in the next two chapters. Beyond this point, the designer must rely on experience, common sense, and imagination.

22.1

cmCUIT

REQUIBEMENTS

The design of any circuit is built upon a statement of requirements. The initially available requirements may range from a broad general account of the purpose of the circuit to a detailed listing of input conditions and the specific actions that must be performed. If the circuit is an entity in itself or belongs to a system which is in the later stages of development, precise requirements are usually available and can be applied directly to the design of the circuit. On the other hand, if the circuit is part of a system whose development has not fully crystallized, it is rarely possible to determine detailed requirements which relate a particular circuit to the others with which it must work. This is particularly true when the design of system components is distributed among several people. The most that can be expected in this situation is a general statement of the job that must be performed by the circuit

498

The Design of Switching

Circuits

and the nature of the input information that will be available. As the designs of the several circuits of the system advance concurrently, the output conditions of one circuit become the input conditions of another, and specific requirements develop as each designer determines his needs and what he can furnish. If the job falls in this category, the designer may have considerable freedom of choice in specifying his own requirements, but suffers from the penalty that his circuit must often be changed to meet new conditions in associated circuits. On the other hand, when requirements can be rigidly stated, there may be a disadvantage of inflexibility in designing a circuit to meet the fixed conditions. Although new and improved methods occur to the designer, they may involve changes in associated circuits that cannot be accepted. The status of requirements, then, varies from job to job. In many cases, the only explicit requirements given to the designer concern the general functions of the circuit. Thereafter the designer must ferret out for himself the mass of detailed requirements dealing with input conditions, limits, output conditions, time relations, necessary precautions, etc. These requirements fall into two categories: those imposed by external conditions over which the designer has little control; and those that develop from the circuit plan, from the form of the circuit itself, and from apparatus and equipment considerations. The first category of requirements must be collected and co-ordinated as completely as possible as the initial step in circuit design. The second category usually appears only during the course of circuit design. When the requirements are known, the job of developing a multifunctional circuit divides into four principle stages, two of which constitute planning and the other two the detailed design. The first stage is the devising of a scheme or method of performing the required functions; the second stage is the embodiment of the scheme in a block diagram form in which the blocks represent interrelated unifunctional units that are known or are realizable; the third stage is the detailed design of the component circuit blocks and their co-ordinating circuit paths to give a schematic representation of the circuit plan. In the final stage, certain practical and commercial considerations must be taken into account involving the addition of apparatus codes, contact numbers, contact protection, etc., to the circuit.

22 .2

UNDERLYING

FACTORS

Before discussing these four stages of design in more detail, several underlying factors which the designer must always keep in mind should be mentioned. Stated in brief, circuits should be designed in such a way that they: (1), meet the functional requirements; (2), are reliable; (3), are economical with respect to manufacture, installation, and life;

General

Principles

of Multifunctional

Circuit

Design

499

and (4), are easy to maintain. The emphasis placed upon the latter three items is, of course, dependent upon the application of the circuit, the conditions of use, and the number of units to be manufactured. The first of these items requires no discussion. The second, reliability. may have a profound effect upon the circuit design. It implies immediately a good basic design with no built-in hazards. Beyond this, the degree of reliability required depends on the job that the circuit has to do and its relative importance in the over-all system. The reliability of a circuit, particularly as it involves checking features added to the basic circuit, must take into account factors of economy and maintenance. A circuit can be rejected from a cost standpoint because of an excess of checking features just as surely as it can be discarded for unreliability. As in so many things, the designer must strike a happy balance. The relationship with maintenance arises from the fact that checking features should be arranged to give adequate indications of trouble conditions in addition to guarding against false operation. This can be taken care of by designing the check circuits so that their manifestations are clear to a maintenance man, and by tying them into appropriate alarm devices and trouble indicators or recorders. This whole matter of reliability is also associated very closely with apparatus. For example, different types of relays are in current use that have single contacts exposed to air, twin contacts exposed to air, and mercury contacts sealed in glass. The probability of an open contact due to dirt decreases rapidly with each successive type. It is obvious that factors such as this should react strongly upon the manner in which a circuit is designed. There are many ways in which the designer can influence the economy of his circuit. The first is in the efficiency with which he makes use of the apparatus. It is axiomatic, of course, that good design is not wasteful of relays, contacts, extra windings, or any of the component apparatus from which circuits are built. However, an attempt for economy should not drive the engineer to such tight design that maintenance expenses in the field may be unduly increased. Choice of types of apparatus is another factor in the cost of circuits. Consistent with reliability and adequate life, the design should be . based on the least expensive apparatus. In relay circuits, for example, the designer should attempt to use high-production relays and avoid special types of relays. In this same respect, apparatus that requires large maintenance effort to keep in adjustment should be used only where absolutely necessary.

In addition to first cost, the factors of service life and maintenance affect the over-all cost of a circuit. It is worthwhile to make original provision for such items as contact-protection networks and

500

The Design of Switching

Circuits

more expensive contact metal, entirely aside from the question of reliability. if the service life can be extended appreciably. However, efforts along these lines must be keyed in with similar accomplishments in the rest of the system. A circuit designed for fifty-year life in a system built for thirty-year life is certainly questionable. A more subtle aspect of circuit design economy has to do with the manufacturability of circuits which are to be produced in quantity. For a given large number of relays, manufacturing costs are less if the over-all circuit can be broken up into small independent units, capable of mass production, than if the circuit must be handled as one big unit. Small units are adapted to assembly line construction, wiring, and testing techniques, whereas large cumbersome frames of relays definitely are not. This implies segregated functional design as opposed to design in which all functions are so inextricably interrelated that it is impossible to divide the circuit into units. From the standpoint of maintainability, if a circuit is designed in a straightforward and logical manner, is reliable, utilizes apparatus simple to adjust and replace, and gives adequate indications of trouble that may occur, it will offer little difficulty in service. 22.3

THE FOUR STAGES OF DESIGN

After this discussion of the background factors of design, the four stages in the development of a multifunctional circuit can be covered more intelligently. The four stages, to repeat, comprise: (1), devising a scheme or method; (2), planning in terms of a functional block diagram; (3), designing the detailed circuit paths and configurations; and (4), applying the practical adjuncts such as contact protection, apparatus codes, and designations. The four stages will be discussed as consecutive steps, although in practice the designer may find it necessary to jump back and forth during the course of development. For example, the first or second stage sometimes depends upon a new circuit concept which may or may not be valid. There is no point in blindly continuing the circuit planning before the new concept has been demonstrated to be feasible by working out at least part of the circuit in some detail. Also it will very frequently happen that new or modified ideas will occur, during the later stages of design, that require changes in the basic plan. Therefore, the following outline should not be construed as a fixed and immutable mode of attack, but rather as a general procedure which can be modified by circumstances. The first stage of design, the devising of a scheme, offers the same difficulties to description as any other creative act. Proficiency depends upon knowledge of the fundamentals of circuit design, familiarity with many so-called unifunctional circuits and their practical

General

Principles

of Multifunctional

Circuit Design

501

applications, and understanding of apparatus capabilities and shortcomings. Superimposed upon all these characteristics, the designer must be able to utilize this knowledge in translating the requirements of a job into circuit terms. Often the basis of a solution will be a simple but new circuit configuration around which can be woven a multiplicity of well known circuit concepts to build it out to complete form. Some times the scheme may involve nothing more than a new relationship among standard functional circuits. Perhaps the designer can merely adapt a previously used circuit to his new requirements. Jn this case, of course, most of the design phase of circuit development is eliminated. The second stage of circuit design, developing a functional plan or block diagram, can be described in more concrete fashion. The process consists primarily of elaborating the basic scheme, in terms of blocks representing circuit functions, to the point where all requirements excepting those that affect only circuit details are satisfied. As far as possible, no detailed design should be attempted at this stage. The designer should draw heavily upon his knowledge of what can be accomplished by simple aggregations of basic circuit elements, such as relays. In particular, he should attempt to think in terms of unifunctional circuits of the type discussed in the preceding chapters. Individual circuit elements such as relays are in themselves too limited in performance to permit broad planning of a circuit. The ultimate objective of this planning stage is a block diagram in which each block represents a single-function circuit, and on which all important interrelationships between blocks are shown. When this objective has been achieved, there exists a framework which vastly simplifies the detailed job of designing the circuit paths and adapting known unifunctional circuits to meet the stated requirements. The alternative, which consists of building the circuit relay by relay, is much too inefficient to be practical. The most satisfactory approach to developing a block diagram is to ·start with a few main subdivisions of the over-all circuit and successively break these down until each block represents a unifunctional circuit. The chief guide during this process is the set of known requirements which must be kept constantly in mind. Jn a surprisingly large number of cases in the planning of switching circuits, familiar functional circuits are found to be applicable. When a new circuit concept is encountered, the designer can usually recognize whether an appropriate circuit can readily be designed. If this is so, the circuit can be designated on the diagram and the design deferred until later. However, in some cases it may be necessary to perform a certain amount of detailed design immediately to insure that the concept is feasible. In some cases, the planning may have to be re-oriented to permit a practical solution.

502

The Design of Switching

Circuits

As the planning progresses, the attempt should be made to obtain the simplest and most efficient arrangement among the various blocks. Duplications of particular functions can often be eliminated by making part of the circuit common, if time relationships permit, or by adding supplementary functions. The same groups of relays can sometimes be used for diverse functions, although this is a course to be handled with caution. Even though apparent economies can be achieved, there is a risk of obscuring the circuit action to the point where understandability and maintenance are adversely affected.

In addition to its value as the basis for detailed design work, the block diagram is an effective means of either disclosing faults in the original plan or suggesting better alternative methods. The value of this can be appreciated when the great amount of time that often is It is obvious that the ability to spent on detailed design is considered. construct effective block diagrams and to think in terms of functional blocks is dependent upon the designer's circuit knowledge. Although this knowledge grows primarily with experience, the designer should from the start make a conscious effort to familiarize himself with different types of basic circuits already in use and to classify them in terms of function. In this way he develops a constantly growing "catalogue" of circuit building blocks which expedites his planning and design of circuits.

In addition to the block diagram itself, it is usually desirable to work out a time or sequence chart of the planned circuit action. Almost inevitably multifunctional circuits operate on a sequential basis, and it is difficult to show this aspect on a static block diagram. The sequence chart is not in terms of individual relay actions, but, rather, it shows the interplay in time among the functional blocks. It should also indicate alternative sequences of operations necessitated either by external requirements or by internal conditions, including troubles encountered by the circuit. Both the block diagram and the sequence chart should include whatever notes are necessary to explain and record their meaning. The block diagram cannot take account of all the detailed requirements that the designer has gathered. Many of these depend upon the specific circuit arrangements that will be used, and some are of such minor importance that there is no need to consider them during the initial planning. However, the final stage of the block diagram should account for all major requirements and should indicate the complete functional breakdown of the proposed circuit. The detailed design work is based largely upon the information contained in the preceding chapters of this volume. Chapters 2 through 10 give the general principles for designing basic circuits utilizing

General

Principles

of Multifunctional

Circuit

Design

503

relays, switches, and electronic elements. Chapters 11 through 21 indicate the technique of designing unifunctional circuits, and present a wide variety of examples of different types of such circuits. For a competent designer, a functional block diagram is a specification which indicates directly what a great deal of the circuit must be like. The detail design job itself breaks down into two major parts. The first of these is the design or choice of circuits to fill the functional blocks, and the second is the integration or co-ordination of these circuit units into a unified whole. The two parts of the job are interdependent and both must be kept in mind at all times. The design of a particular functional unit is affected by its relationship with all other functional units and by the manner in which they are associated. In the same sense, the association must be developed in accordance with the characteristics of the individual units. In any case, both jobs represent an extended application of the basic techniques for design of circuit control paths. The starting point of the design will normally be the selection or design of circuits to fulfill the individual functions. In general, work should start on the more difficult or unfamiliar functions, since the solution to them may modify the rest of the circuit. As the design progresses and additional units take form, the interconnecting and coordinating paths between circuit units can be developed. For some of the functional blocks the designer is often faced with the decision of either designing a new circuit unit or adopting a previously used one. It is normally a temptation to follow the former course on the basis that thereby the needs of the particular circuit can most precisely be met. Pride also enters into this. However, there are definite advantages in adapting more or less standard circuit configurations to new circuits, when possible. A great deal of time may be saved that otherwise would be spent in a perhaps fruitless search for a new circuit arrangement. In general, such circuits have been proven under fire and their reliability is known. In many cases, there may be the benefit from a maintenance standpoint of some degree of standardization in the field. The foregoing is not intended to limit the designer to copying old methods. The designer should al ways aim for improved and more reliable circuit configurations if he has reason to believe that his objective can be attained. However, if this justification is lacking, and if requirements do not force him in a new direction, adaptation of previously designed functional circuits is the preferable course. The detailed design of circuits will often disclose situations where it is possible to overlap or intermingle functions. This may appear economical of apparatus, yet it can be carried to a point where over-all economy suffers. The danger point is approached when it

504

The Design of Switching Circuits

becomes difficult to understand the circuit, or to recognize what functions a given part of the circuit is performing. This is of particular importance when people other than the designer must build and maintain the circuits, since extra time spent in testing and maintenance may cost more over a period of time than the original price of a few additional relays. However, these considerations are modified by the intended use of the circuit. Circuits which are to be manufactured in large quantity are very sensitive to cost. Furthermore, circuits of this type are usually relatively simple in nature, and the doubling up of functions does not overcomplicate their understanding. On the other hand, with larger and more intricate circuits such as the major components of computer or telephone switching systems, more emphasis should be placed on straightforward design. This should, of course, not be construed as an excuse for wasteful design in these large circuits. During the detailed design, the designer must continually check his circuit against all the original requirements plus all the additional requirements that inevitably develop from the particular circuit method he is using. All conceivable combinations of circumstances which might affect the circuit operation must be considered. Such matters as release under all normal and trouble conditions, guarding against attempted use when in trouble, reaction to trouble in other circuits, etc., are important. The degree and type of checking and trouble-indicating features, as discussed previously, should be carefully worked out. Finally, all limiting, marginal, and race conditions must be investigated to insure that adequate working margins are available. When the circuit has reached the stage on the drawing board where it theoretically meets the requirements, there are usually many final engineering details to be worked out. Although it is frequently necessary during the detailed design stage to choose particular apparatus for parts of the circuit, much apparatus will be of a generalpurpose nature where the code selection can be deferred. A major part of the final stage of circuit design, then, consists of making the choice of apparatus to fit the particular needs of the circuit. Among the factors to be considered are speed of operation, contact metal on relays, voltage and wattage limits, tolerance ratings, and magnetic, electrostatic, and vibration interference with certain types of apparatus. The designer is often required to engineer circuits for some minimum life, perhaps ten, twenty, or forty years. This necessitates a calculation of the number of operations of such apparatus as relays, and the provision of adequate contact protection. Another item that is normally deferred until the final stage of design is the fusing of the circuit. This should be arranged so that there is adequate protection against the dangerous effects of trouble-crosses and short-circuits, and yet unnecessary complications should not be

General

Principles

of Multifunctional

Circuit

Design

505

introduced into the circuit wiring. Where the circuit is part of a large system, the fusing should be arranged, if possible, to minimize the disruptive effects of a single fuse failure upon the functioning of other parts of the system. An associated aspect of the general design ptoblem is laboratory testing. When commercial production is involved, it is usually desirable to set up a test model of a large, complicated, or critically important circuit. Rigorous testing of the model under the worst probable operating conditions often discloses defects or errors in the original circuit which should be changed before production. Testing may also indicate the need for improvements in operating conditions or desirable changes in requirements. The discussion in this chapter has been intended to highlight the general principles of multifunctional circuit design. It is obvious that the major portion of circuit design involves creative effort, for which it is al ways difficult to establish specific rules and procedures. For this reason, the present discussion will be implemented by an illustrative example of the correct approach to a design problem and a typical circuit solution, to be given in the next two chapters. The primary emphasis is upon the creative aspects of design. That is, the text will deal with the development of theoretical circuits in the sense that particular apparatus will not be specified, and little or no mention will be made of what has been designated the fourth stage of design.

Chapter 23 PLANNING A MULTIFUNCTIONAL CIRCUIT

In this chapter and the next, the methods outlined in Chapter 22 will be applied to a specific design problem to illustrate both the method and the underlying principles. The treatment of the problem is divided into two main parts: the planning of the circuit, covered in this chapter; and the detailed design work, discussed in the next chapter. The planning phase of the job will include: determining all the requirements of the circuit; laying out general and tentative block diagram solutions to the problem; and finally breaking down the general block diagram to specific functional form which satisfies the general and detailed requirements. In the final form, each block will represent as far as possible a clearly definable circuit function which can be treated with reasonable independence during the detailed design stage. As was mentioned in the preceding chapter, the design method is very much subject to the individuality and experience of the designer. Given the same design problem, two designers will often plan along completely different lines and produce equally good functional block diagrams. Thus, in the attack on the problem as presented in this chapter, the various steps taken are not necessarily the only logical ones. The particular approach, rather, is one which will best illustrate the method in general. 23.1

STATEMENT OF THE PROBLEM

The example chosen for illustration is part of an automatic telephone switching system and is called a "Dial Pulse Register Circuit". This circuit is well defined and, except for a simple input connection and one set of output conditions, does not depend upon other circuits of the system for its operation. The first step in the approach to the problem is to express the purpose of the circuit. In this case, the purpose can be broadly summarized as follows: to receive pulse signals representing four digits from a telephone dial or similar calling device; to keep a record of the digits; and, when the fourth digit has been received, to make the numerical information available in a form suitable for use by an external circuit. - 506 -

Planning

a Multifunctional

Circuit

507

Before an exact statement of the requirements to be met by the circuit can be made, the conditions under which it must operate and the actions which it must perform must be determined. First consider the input conditions which convey information to the circuit. The input to the register circuit is delivered over a pair of wires which is connected to a telephone line. In the telephone system this is not a permanent connection, but is established through connector circuits at the time the subscriber wishes to place a call. The design of these circuits is not a part of the present problem, and their exact nature is unimportant since the register circuit does not function witil after this connection is made. At the subscriber's end of the telephone line is a telephone set which contains a dial. The telephone set is arranged so that, when the handset is in the cradle, an open circuit exists between the pair of line wires. When the handset is lifted, contacts operate to place the voice transmission equipment in the circuit in such a manner as to form a d-c current path between the two line wires. Contacts on the dial, which are normally closed, are include\' U'\t!'